A volcano does not set an alarm
An eruption can begin after days of conspicuous unrest, after years of intermittent signals or with little warning from a poorly monitored volcano. Volcanologists cannot see a neat tank of magma beneath the summit and read a pressure gauge. They infer what is moving underground from the volcano's response at the surface.
That response can be remarkably subtle. A slope swells by centimetres. Small earthquakes cluster or migrate. The mix of escaping gases changes. A patch of ground grows warmer. None of those observations alone means that lava will soon appear. Together, and compared with the volcano's own usual behaviour, they can show that its plumbing is changing.¹
Listening to rock break
Magma pushing through rock can fracture it, producing earthquakes. A local network of seismometers records where tremors occur, how deep they are and whether the pattern shifts. A swarm moving upwards may suggest a moving magma body; continuous tremor can accompany fluids flowing through the system. But volcanoes can also quake without erupting, and ordinary tectonic earthquakes can occur nearby. Interpretation depends on the place, its history and other measurements.
An important distinction is between the count of earthquakes and their pattern. Ten small tremors beneath a busy volcanic system may be normal. The same number at a usually quiet volcano, becoming shallower and accompanied by swelling, deserves more attention. Scientists therefore build a baseline over months and years. Monitoring is most useful before a crisis begins. At some volcanoes, instruments are sparse or damaged by harsh weather, so scientists also use satellite observations and reports from people nearby. The absence of a detected signal is less reassuring when the network has blind spots. A forecast should therefore make clear how well the volcano is actually being watched.
Measuring a mountain's breathing
As magma accumulates or moves, it can press on surrounding rock. Ground-based satellite-navigation receivers can detect slow shifts in position; tiltmeters measure changes in slope; radar images from orbit can map broad inflation or deflation. The US Geological Survey uses these methods to track magma pathways and pressure changes.²
Imagine inflating a balloon beneath a thick rug. The rug rises, but its shape depends on the balloon's depth and the rug's stiffness. A volcano is much more complicated: faults, old magma bodies and heated water can all influence surface motion. Scientists fit models to deformation measurements, then test whether earthquakes and gases point to the same explanation. Rising ground is evidence of change, not a guaranteed eruption clock.
What escapes before magma?
Volcanic magma contains dissolved gases. As it rises and pressure falls, gases can escape and reach the surface before an eruption. Teams measure emissions such as sulphur dioxide and carbon dioxide, using instruments on the ground or remote observations. Changes in emission rate or gas ratios may indicate that fresh magma is moving or degassing. The USGS notes that gas and water chemistry can reveal changes in the magma system.³
Gas is tricky because groundwater and the volcano's own conduits can filter or divert it. A low reading may mean little fresh gas, or it may mean gas is not reaching the sensor. Some volcanoes emit hazardous gases while otherwise appearing quiet. Measurements must be made safely and interpreted alongside geology, weather and the volcano's previous behaviour.
Heat offers another clue. Thermal cameras and satellites can detect hotter ground, vents or water. A newly warm area may indicate rising magma, but hot fluids can circulate without an eruption. Weather and vegetation can obscure a signal. Again, the trend and its agreement with other observations matter more than a single bright pixel.
Forecasting is a changing judgement
Scientists use monitoring data and the record of past eruptions to build possible scenarios: an intrusion that stalls, a small eruption, a larger explosive event, a lava flow in one direction rather than another. The USGS says probabilistic approaches help express the uncertainty in such forecasts.⁴ A forecast can improve as new evidence arrives, and it must be revised if the volcano behaves differently.
The hard questions are practical. How soon might an eruption start? Where might a vent open? Will ash threaten aircraft or communities? Can rain turn loose ash into dangerous mudflows? Different hazards demand different actions. A volcano may pose limited danger to people on the ground while its ash plume makes flying unsafe, or vice versa.
For that reason, US volcano notices use separate alert terms for ground hazards and colour codes for aviation. “Watch” signals heightened or escalating unrest with an uncertain time frame; it is not a promise that an eruption will occur on a particular day. “Warning” indicates a hazardous eruption is imminent, under way or suspected.⁵ Other countries use their own systems, but the communication challenge is shared.
A good warning explains both what has changed and what remains uncertain. Authorities may need to close an area before the evidence is conclusive, because waiting for certainty could remove the chance to leave safely. Conversely, an intrusion can stop without erupting. Treating every uneventful warning as a failure would discourage the very caution that makes monitoring valuable.
The science is strongest when its limits are plain. Volcanologists can often recognise unrest and narrow plausible outcomes; they cannot guarantee the precise minute, size or location of an eruption. The mountain is not silent, but it speaks in several imperfect instruments at once.
