Volcanologists rely on a broad and constantly evolving toolkit that spans everything from seismometers buried in the ground to satellites orbiting hundreds of kilometers overhead. No single instrument can capture the full behavior of a volcano, so modern monitoring weaves together seismic, geodetic, chemical, thermal, acoustic, and computational methods to build a picture of what is happening beneath and above the surface. The combination of these tools is what makes eruption forecasting possible, even though no forecast is ever a sure thing.
Seismic Monitoring
Seismometers are the oldest and most fundamental instruments in a volcanologist’s arsenal. Magma forcing its way through rock, gas bubbles collapsing in underground conduits, and fractures opening under pressure all generate seismic waves that seismometers can detect. A network of sensors spread around a volcano lets researchers triangulate where these signals originate and track whether the source is getting shallower over time, which is one of the strongest clues that an eruption may be approaching.
One particularly revealing seismic signal is harmonic tremor, a sustained, rhythmic vibration that typically indicates magma or fluids are moving through narrow channels underground. Because tremor can last for minutes to hours, it stands out against the background noise of ordinary earthquakes. During a period of heightened activity at a restless volcano, automated algorithms have detected over a thousand minutes of harmonic tremor, and these episodes generally preceded spikes in explosive activity.1Geophysical Research Letters. Automated detection and characterization of harmonic tremor in continuous seismic data Recognizing tremor patterns in real time helps observatory staff judge whether a volcano’s plumbing system is actively pressurizing.
A newer twist on seismic monitoring is distributed acoustic sensing, or DAS, which converts a standard fiber-optic cable into thousands of closely spaced vibration sensors. A deep-learning system tested on DAS data recorded during the 2021 La Palma eruption was able to detect and classify volcano-tectonic events with roughly 97 percent accuracy while tracking how each event evolved over time.2Journal of Geophysical Research: Solid Earth. RNN‐DAS: A New Deep Learning Approach for Detection and Real‐Time Monitoring of Volcano‐Tectonic Events Using Distributed Acoustic Sensing Because fiber-optic cables can be laid along the seafloor or through areas where conventional seismometer stations would be impractical, DAS opens up monitoring in places that were previously blind spots.
Measuring Ground Deformation
When magma accumulates underground, it pushes the surrounding rock outward and upward, inflating the surface like a slow-motion balloon. When magma drains away, the surface sinks. Tracking these centimeter-scale shifts is one of the most direct ways to infer what is happening at depth, and volcanologists have several tools for the job.
GPS receivers fixed to the flanks of a volcano continuously record their three-dimensional position. Paired with tiltmeters, which measure tiny changes in slope, and strainmeters, which detect how rock is being squeezed or stretched, these ground-based instruments capture deformation in real time. A comprehensive approach that combines GPS surveys, tiltmeters, strainmeters, and microgravity surveys can reveal both the spatial pattern and the timing of ground movement across an entire eruption cycle.3Reviews of Geophysics. A comprehensive approach to monitoring volcano deformation as a window on the eruption cycle
Satellite radar adds a wide-area view. Interferometric Synthetic Aperture Radar, commonly called InSAR, bounces radar pulses off the ground from orbit and compares images taken at different times to map surface displacement across broad swaths of terrain. High-resolution SAR imagery from satellites has proven capable of detecting uplift inside a volcanic crater before eruptions begin. At Agung volcano in Indonesia, researchers used data from multiple satellite missions to spot localized deformation inside the summit crater in the weeks leading up to the November 2017 eruption.4Journal of Geophysical Research: Solid Earth. High‐Resolution InSAR Reveals Localized Pre‐Eruptive Deformation Inside the Crater of Agung Volcano, Indonesia The great advantage of InSAR is that it works on volcanoes where no ground instruments have been installed, which is the case for most of Earth’s roughly 1,500 potentially active volcanoes.
Volcanic Gas Monitoring
Magma contains dissolved gases, mainly water vapor, carbon dioxide, and sulfur dioxide, that escape as magma rises toward the surface and pressure drops. Changes in the amount and composition of volcanic gas provide a chemical window into what the magma is doing. A sudden spike in sulfur dioxide emissions, for example, often signals that fresh magma has arrived at shallow depths.
One ground-based technique measures the invisible stream of carbon dioxide seeping through volcanic soil. An accumulation chamber placed on the surface traps the gas as it rises, and a sensor inside tracks how fast the CO₂ concentration builds. That rate, combined with the known volume and footprint of the chamber, gives a flux measurement at that specific point.5Journal of Volcanology and Geothermal Research. A low-cost IoT-based accumulation chamber for measuring diffuse CO2 fluxes Survey teams walk these chambers across a volcanic field point by point to map out where gas is leaking most intensely, helping identify hidden faults or rising magma bodies that have no surface expression otherwise.
Drones have transformed gas sampling in the past decade. Rather than sending people into the toxic, superheated plume billowing from a crater, volcanologists now fly lightweight uncrewed aircraft directly through it carrying miniature gas sensors. During flights at an active volcano, one drone-mounted sensor system measured the ratio of CO₂ to SO₂ in the plume at roughly 28 to 33, with the measurements closely matching those from a separate verification sensor carried on the same platform.6Scientific Reports. Observing volcanoes with drones: studies of volcanic plume chemistry with ultralight sensor systems The CO₂-to-SO₂ ratio matters because it shifts as magma rises or stalls: a changing ratio can tell scientists whether new magma is ascending or the system is settling down.
A deeper kind of chemical analysis happens back in the lab. Volcanologists collect rock and lava samples, then study tiny pockets of melt trapped inside growing crystals, called melt inclusions. The water and CO₂ dissolved in these inclusions record the pressure at which the crystal grew, which translates directly into depth. At one volcano, melt inclusions revealed that crystals had formed at depths ranging from about one kilometer to nearly nine kilometers below the summit, allowing researchers to reconstruct how magma migrated upward before the eruption.7PubMed Central. Crystal and melt inclusion timescales reveal the evolution of magma migration before eruption This kind of forensic petrological work does not provide real-time warnings, but it fills in the underground story after the fact and helps calibrate what the real-time instruments are detecting.
Thermal Cameras and Satellite Hotspot Detection
Volcanic activity almost always changes the surface temperature, whether through lava flows, fumaroles venting steam, or heated ground. Thermal infrared cameras capture these heat signatures and have become standard equipment at well-monitored volcanoes. They can be hand-held for field surveys, mounted on tripods or helicopters, or installed permanently for continuous surveillance. Their data feed into estimates of surface temperatures, lava effusion rates, and heat and mass fluxes, all of which help volcanologists judge how vigorously an eruption is proceeding.8Earth-Science Reviews. Volcano surveillance using infrared cameras
From orbit, satellites offer global coverage. A recently developed algorithm processes scenes from the VIIRS satellite instrument, which offers a spatial resolution of 375 meters and multiple passes per day, to flag thermal anomalies at volcanoes worldwide. It can pick up temperature increases as small as half a degree above background while keeping the false positive rate under about two percent.9Remote Sensing of Environment. TIRVolcH: Thermal Infrared Recognition of Volcanic Hotspots That sensitivity matters because a subtle thermal brightening at a crater lake or fumarole field can be the earliest satellite-detectable sign that something is changing underground. Satellite thermal monitoring is especially valuable for the many remote volcanoes that have no ground-based instruments at all.
Listening for Eruptions with Infrasound
Explosive eruptions push a massive pulse of low-frequency sound into the atmosphere. These infrasonic waves, below the threshold of human hearing, travel enormous distances because they lose energy slowly in the atmosphere. Infrasound is the only ground-based monitoring technique capable of detecting explosive eruptions from thousands of kilometers away. Researchers have demonstrated that infrasound array analysis, using both the acoustic amplitude and how persistently the signal is detected, can automatically identify eruptions of Mount Etna from stations more than 500 kilometers away in near-real time.10Scientific Reports. Long range infrasound monitoring of Etna volcano
Because infrasound couples directly to the explosive process itself, it has potential as an early warning tool. A dedicated processing method demonstrated that infrasound arrays can detect the onset of an explosive eruption and automatically notify authorities, relying on the physical link between the explosion and the sound wave rather than on probabilistic forecasting.11Journal of Geophysical Research: Solid Earth. Infrasonic Early Warning System for Explosive Eruptions For volcanoes near populated areas or busy flight corridors, shaving even a few minutes off the notification time can matter for evacuation decisions and aviation safety.
Peering Inside with Geophysics
Seismometers and GPS tell you that something is moving or swelling underground, but they do not directly image what the interior looks like. Several geophysical techniques attempt to fill that gap.
Muography borrows a concept from particle physics. Cosmic rays constantly shower the atmosphere and produce muons, heavy cousins of electrons that can penetrate hundreds of meters of rock. By placing a muon detector on a volcano’s flank and measuring how many muons arrive from different directions, researchers can map the density structure inside the edifice, because dense rock absorbs more muons than porous or molten material. Although muography allows direct density measurements inside a volcano, the technique has limitations: scattered particles following indirect paths can cause the measured density to come out lower than the true value.12PubMed Central. Muographic imaging with a multi-layered telescope and its application to the study of the subsurface structure of a volcano Still, it has been used to locate dense plugs of solidified lava inside conduits and to identify zones of weakened rock prone to collapse.
Magnetotellurics uses naturally occurring electromagnetic fields to probe electrical resistivity deep underground. Molten rock and hot brines conduct electricity much better than solid, cold rock, so regions of low resistivity often indicate magma storage zones. Beneath Volcán Uturuncu in Bolivia, broadband magnetotelluric data revealed a zone of very low resistivity at about 15 kilometers depth, interpreted as a large body of andesitic melt requiring more than 20 percent melt fraction. Shallower conductive bodies around five kilometers depth were attributed to a combination of partial melt and circulating fluids.13Geology. Magnetotelluric images of magma distribution beneath Volcán Uturuncu, Bolivia
Microgravity surveys offer another angle. A precise gravimeter measures tiny changes in the pull of gravity at the surface, which shift when mass moves underground, as when magma intrudes into new space or retreats. While the technique has been used for decades to understand volcanic plumbing, researchers argue it has untapped potential as an eruption forecasting tool. Gravity changes may detect long-term mass buildup before the vigorous seismicity and deformation that typically trigger alerts have even begun.14Journal of Volcanology and Geothermal Research. Microgravity as a tool for eruption forecasting That possibility is appealing because it would extend the forecasting window backward in time, catching a volcano in an earlier stage of unrest.
Monitoring Volcanoes Underwater
Most of Earth’s volcanic activity happens on the ocean floor, far from the eyes of satellite sensors and beyond the reach of most conventional instruments. Submarine monitoring is difficult and expensive, but hydrophones, essentially underwater microphones, can pick up the acoustic signatures of seafloor eruptions from considerable distances.
During a five-month deployment in the Manus Basin of the Bismarck Sea, a deep-sea hydrophone system recorded evidence of a gas-driven explosive eruption that lasted roughly 60 days. The eruption raised local sound levels by about 12 decibels on average, a substantial increase. The data also showed that eruptive episodes could start and stop abruptly and persist for months.15Geophysical Journal International. Acoustic evidence of a long-lived gas-driven submarine volcanic eruption in the Bismarck Sea Discoveries like this highlight how much volcanic activity goes unnoticed simply because we lack permanent monitoring infrastructure on the seafloor. Advances in ocean-bottom sensor networks and the repurposing of fiber-optic cables for DAS could help close this gap over the coming decades.
Computational Models and Alert Systems
Raw data from seismometers, GPS stations, gas sensors, and satellites are only useful if someone, or something, synthesizes them quickly enough to inform decisions. Increasingly, that synthesis is handled by automated systems.
At Mount Etna, one of the most heavily instrumented volcanoes in the world, researchers built a probabilistic model that ingests measurements from multiple monitoring streams and estimates the probability that the volcano is in a given state, whether quiet, showing signs of unrest, or actively erupting. The system encodes the relationships between different types of observations so that conflicting or ambiguous data can be weighed against each other, giving on-duty personnel a continuously updated picture.16Journal of Geophysical Research: Solid Earth. A multivariate probabilistic graphical model for real‐time volcano monitoring on Mount Etna
On the hazard-response side, ash dispersion models forecast where volcanic ash clouds will drift after an eruption, which matters for aviation and communities downwind. In New Zealand, a jointly developed system feeds eruption parameters and high-resolution weather data into the HYSPLIT dispersion model, a widely used tool that tracks how particles are carried by wind, to produce probabilistic ashfall forecasts in near-real time.17Journal of Applied Volcanology. Towards real-time probabilistic ash deposition forecasting for New Zealand These forecasts help emergency managers decide when to issue warnings, reroute aircraft, and prepare for ash cleanup. The probabilistic approach is important because wind conditions and eruption intensity are both uncertain; rather than producing a single predicted ash footprint, the model generates a range of scenarios weighted by likelihood.
Reading Eruption History in Tree Rings
Not all volcanological tools are electronic. Understanding how often a volcano has erupted in the past, and how big those eruptions were, is essential for estimating future risk. Volcanologists piece together eruption histories from geological deposits, ice cores, and even tree rings.
Large explosive eruptions inject sulfate aerosols into the stratosphere, dimming sunlight and cooling temperatures regionally or globally. Trees at high-elevation forest borders record this cooling as unusually narrow growth rings. A study using bristlecone pine chronologies spanning 5,000 years found that over the last millennium, years of minimum ring width could be matched with known eruptions or volcanic signals in ice cores in 86 percent of cases. Growth dropped significantly in the year following an eruption and stayed suppressed, though less dramatically, for several years afterward.18Quaternary Research. Bristlecone pine tree rings and volcanic eruptions over the last 5000 yr Tree-ring records help fill in the eruption timeline for periods and regions where written records do not exist, giving volcanologists a longer baseline against which to judge whether a particular volcano’s current behavior is unusual.
Why No Single Tool Is Enough
Every monitoring method has blind spots. Seismometers miss eruptions that produce little seismicity. InSAR struggles in vegetated tropical terrain where radar signals scatter. Gas measurements depend on weather conditions that can dilute or redirect a plume. Infrasound detects explosions beautifully but tells you nothing about a quiet effusive lava flow. Thermal satellites revisit the same spot only a few times a day, meaning a short-lived explosion can slip between images. Muography requires months of exposure and can only view a volcano from limited angles.
The strength of modern volcanology lies in redundancy and integration. When seismicity rises and GPS stations on the same flank start moving apart and gas emissions jump simultaneously, the converging evidence is far more convincing than any one signal alone. The push toward automated systems that fuse these streams is partly about speed, getting alerts out faster, and partly about catching the eruptions that any single tool would miss. It is also about extending coverage to the hundreds of volcanoes worldwide that currently have little or no monitoring at all, where satellite-based tools and low-cost sensor packages could provide at least a baseline level of watchfulness.