Researchers are hailing a breakthrough in the study of volcanic eruptions, as new ways of modeling data might make this science as reliable as a weather forecast.
Scientists from Institut des Sciences de la Terre (ISTerre) in France have been the first to predict the behavior of a volcano with data assimilation techniques, which up to now were mainly used to predict weather.
Grace Bato, lead author of the paper published in the journal "Frontiers of Earth Science," said that with this technique they would be able to give daily or hourly predictions on volcanic activity.
Data assimilation uses satellite measurements and mathematical calculations to make predictions based on the effect that moving magma has on the earth's surface.
The satellites use radar and GPS to measure the deformation of the earth's surface as magma moves underneath.
Similar to the way satellite information is used in meteorology, the volcano researches use the data to create models of how the magma is moving and then run these models into the future.
Researchers have previously used these techniques to predict the effect of carbon dioxide emissions on the earth's climate and weather.
The volcano researchers were helped by the large quantity of GPS data that is now available thanks to the increased use of GPS systems on the ground.
Lead author Bato and her colleagues began by creating a computer model of a particular kind of volcano, which sits on top of a build-up of magma but does not erupt very explosively.
The researchers' model allowed them to predict how much pressure is needed for these kinds of volcano to erupt.
Bato and her colleagues also predicted the shape of the magma reservoir under the volcano and the rate of magma flow into the reservoir - features that are very difficult to measure because of their depth.
The researchers have begun testing their method on the Grímsvötn Volcano in Iceland and the Okmok Volcano in Alaska, in the hope that more accurate predictions could save many lives in the future.