How Volcanoes Are Monitored: Seismic Networks, Gas Sensors, and Ground Deformation

Monitoring is a three-instrument problem, not a single sensor

No single instrument can reliably forecast a volcanic eruption. Observatories instead fuse three largely independent data streams — seismicity, gas geochemistry, and ground deformation — because each responds to a different physical process and each has different failure modes and blind spots. Seismic networks detect the mechanical signature of magma or fluid movement through rock; gas sensors detect the chemical signature of magma degassing as it rises and depressurizes; deformation instruments detect the physical signature of a magma reservoir inflating or deflating. A volcano can show a strong signal in one of these channels while staying quiet in the others, and the interpretation differs sharply depending on which combination is observed — which is why modern eruption forecasting is fundamentally a data-fusion exercise, not a threshold check on any one sensor.

Seismic networks: reading the mechanical signal of moving magma

Volcano-monitoring seismometer networks track two things that matter more than simple earthquake count: the depth trend of events over time, and the event type. A swarm of small earthquakes migrating from several kilometres depth toward the surface over days to weeks is a classic signature of magma forcing its way upward through the crust, fracturing rock as it goes (volcano-tectonic, or VT, earthquakes). A separate and often more diagnostic class is long-period (LP) events — lower-frequency signals associated with resonance in fluid-filled cracks and conduits, generally interpreted as fluid or gas movement rather than pure rock fracture, and often considered a stronger precursor to eruption than VT swarms alone because they more directly implicate magma or volatile movement rather than generic tectonic stress relief.

A practical field screening index combines daily event count, average local magnitude, and average focal depth into a single number, with shallower average depth weighted to increase the index — since a shift of seismicity toward the surface indicates magma is approaching the vent rather than stalling at depth. As a planning threshold: a roughly three-to-fivefold increase in daily event count over the local background rate, especially combined with shallowing depth, is a common trigger for elevating an observatory's alert level. Networks rely on automated event detection algorithms for real-time triage, with human analyst review to filter out noise from wind, ocean surf, or regional tectonic earthquakes unrelated to the volcano — automated systems alone are prone to false triggers in noisy environments, so the human-in-the-loop step remains standard practice at most observatories.

Gas geochemistry: SO2 flux and the CO2/SO2 ratio as a chemical barometer

Magma releases dissolved gases as it rises and pressure drops, and the composition of that gas is a genuinely distinct signal from seismicity because it responds to depressurization chemistry rather than mechanical rock failure. Sulfur dioxide (SO2) flux — the mass rate of SO2 emission, typically measured in tonnes per day — is the most widely tracked gas because it is chemically inert enough in the atmosphere to be measured reliably at a distance, and its emission rate correlates with the volume of fresh magma degassing near the surface. Ground-based instruments like the DOAS (Differential Optical Absorption Spectroscopy) network measure SO2 by analyzing how the gas absorbs specific ultraviolet wavelengths in scattered sunlight, while satellite instruments like TROPOMI extend that same absorption-spectroscopy principle to a global, twice-daily monitoring capability for major eruptions and persistently degassing volcanoes.

The ratio of CO2 to SO2 in emitted gas is arguably a more diagnostic single number than either gas alone, because CO2 exsolves from magma at greater depth than SO2 (it's less soluble in silicate melt), so a rising CO2/SO2 ratio is generally interpreted as gas ascending from deeper, less-degassed magma — a signal of fresh magma input into the system rather than continued outgassing of magma that's already been sitting near the surface. Hydrogen sulfide (H2S) is tracked as a faster-responding but more locally variable indicator, since it reacts quickly with atmospheric oxygen and is strongly influenced by the specific hydrothermal plumbing of individual fumaroles, making it more useful as a local early-warning signal at specific vents than as a system-wide index. Because wind disperses gas plumes unevenly, gas flux measurements are routinely paired with meteorological data to separate real emission-rate changes from apparent changes caused by shifting wind patterns, and drone-mounted gas sensors are increasingly used to sample directly over active craters or in terrain too hazardous for ground crews.

Ground deformation: GPS, tiltmeters, and InSAR

As a magma reservoir inflates with new magma or pressurizes from trapped gas, the ground surface above and around it physically deforms — usually uplifting and expanding horizontally, though deflation and subsidence are equally informative when they follow an eruption as the reservoir empties and depressurizes. Three complementary instrument types capture this: continuous GPS/GNSS stations measure absolute 3D ground position with millimetre-level precision over time, tiltmeters measure minute changes in ground slope (in microradians) at a single point with very high sensitivity to short-term changes, and InSAR (Interferometric Synthetic Aperture Radar) uses satellite radar to measure ground displacement over wide areas by comparing the phase of radar signals reflected from the same patch of ground on different satellite passes.

Each method has a distinct strength: GPS networks give continuous real-time point measurements ideal for tracking an evolving signal hour by hour; tiltmeters are extremely sensitive to short-term, localized changes that can precede small eruptive events by hours; and InSAR's key advantage is spatial coverage — it can map deformation across an entire volcanic edifice, including inaccessible terrain with no ground instruments at all, though typically at a slower revisit cadence (days) than continuous GPS. A composite deformation index weighting uplift rate, tilt change, and GPS network expansion gives observatories a single number to track escalation, but the critical caveat repeated in the volcanological literature is that deformation alone is ambiguous — sustained inflation without accompanying seismic or gas escalation can simply stall and never lead to eruption, so deformation trends are interpreted in combination with the other two data streams rather than as a standalone trigger.

From data streams to alert levels and public communication

Observatories convert this fused, multi-instrument picture into standardized public alert levels (commonly a four-tier colour-coded scale: green/normal, yellow/advisory, orange/watch, red/warning) that map to specific, pre-agreed actions — from continued routine monitoring at green, through community briefings and evacuation-route preparation at orange, to active evacuation orders at red. Instrument calibration and network maintenance schedules matter more than they might appear to from the outside: seismometers are typically recalibrated quarterly, gas analyzers monthly (since sensor drift directly biases flux estimates), and GPS stations at every field service visit, because a subtly miscalibrated instrument doesn't fail loudly — it just quietly corrupts the baseline that everything else gets compared against.

A separate but critical communication channel runs to aviation authorities through Volcanic Ash Advisory Centers (VAACs), since ash plumes are a severe hazard to jet engines and observatories maintain agreements to relay plume height and drift-direction updates on a fixed schedule (commonly every six hours) during an active eruption. On the ground, clear colour-coded public messaging, pre-distributed evacuation maps, and regular community drills are treated as being just as essential to reducing casualties as the instrument network itself — a technically excellent monitoring system that the public doesn't trust or understand still fails at its actual job of keeping people safe.