Volcano analysis from space in real time

Humans have always wondered: Can a volcanic eruption be predicted? Sometimes yes! My research group is working on a satellite-based solution for this.

Glowing magma is flung into the air, lava flows from crevasses: For the first time in 50 years, a volcano erupted on the Canary Island of La Palma. The mountain that suddenly erupted doesn’t even have a name. Thousands of people suddenly had to be brought to safety. Experts had said that it was impossible to know if it will erupt in a few days or in a few weeks or never.

Are we really so helpless and at the mercy of a force of nature like a volcano?

The average deformation over the Colima volcano, Mexico, for the 2017-2019 time period using German TerraSAR-X data (Image Credit: TerraSAR-X, License: Creative Commons)

Explosive volcanic eruptions can often announce themselves: The dynamics of gas and magma flows inside the mountain change noticeably before eruptions and cause, among other things, the uplift and subsidence of the volcano’s surface, which is recorded by satellites.

To better analyze and interpret such changes, my colleagues and I at the German Research Centre for Geosciences have developed a data oriented has been developed to detect, millimeter-scale surface deformations automatically, using satellite measurements.

The results achieved using this method provide an important basis for analyzing the growing amounts of satellite data on more than 1,500 active volcanoes worldwide virtually in near real time, and thus ultimately providing more precise warnings of eruptions.

How to detect and interpret the signs of impending volcanic eruptions early and precisely is the primary goal of our research. The assessment of volcanic activity is determined by various parameters, which include seismic measurements, observations of temperature, the composition of released gases – and the often very complex surface deformation. This surface deformation is the focus of the presented research.

The breathing of volcanoes

Even in Greek mythology, there are analogies between volcanic activity and human breathing. And indeed: the surface of many volcanoes rises and falls measurably, almost as if they were breathing in before they spew out gas, ash or lava. Such uplift and subsidences are now recorded by the satellites orbiting the Earth. With the help of Synthetic Aperture Radar technologies, it is possible to get imagery from the Earth’s surface in all weather conditions. Each time the satellites orbit the Earth, one can compare the reflected microwaves with the previous one and use it to calculate an interference pattern that reflects the changes in the ground.

Ground surface deformations occurring at volcanoes, however, are often only in the order a few millimeters to centimeters. In the satellite recordings, they are superimposed by fluctuations in physical characteristics of the ground surface or by atmospheric artefacts. Our newly developed method of computer-assisted data analysis brings significant progress in the interpretation of satellite images.

New machine learning approaches

Until now, satellite images had to be viewed and evaluated manually by scientists. In particular, surface changes that last only for a short period of time are often less examined, although they can provide important information about the inner processes of a volcano. In order to decipher the superimposed signals in the satellite images, scientists around the world have already used artificial intelligence in recent years.

Our method has now adopted a new approach to optimize previous ML algorithms. This approach is based on the application of the Independent Component Analysis (ICA), which tries to extract the latent deformation signals from the satellite measurements. The minimum spanning tree-based approach then compares the multiple iterations of this ICA algorithm and filters out those signals that are most likely to indicate actual surface displacements based on statistical significance.

This method has been tested on several data sets, including satellite images of the Volcan de Colima in western Mexico and Mt. Thorbjorn in Iceland. This algorithm then detected several episodes of previously unnoticed deformation events. It is possible that volcanoes rise and fall even more frequently than was previously known. In these various case studies, it was possible to detect signs of both newly occurring deformations and changes in ongoing deformation processes.

This study shows that the computer-assisted evaluation of satellite images using this new algorithm detects episodes of surface deformation much more precisely and reliably than before. Such a quasi-automated procedure is urgently needed to evaluate the constantly growing amount of observation data from the approximately 1500 active volcanoes.

Real-time analysis possible

The primary goal of our work is to observe volcanic activity in near real time. The satellite data are freely available. Highly precise and reliable algorithms like the one we have developed could evaluate the images even directly in cloud computing platforms. This would eliminate the time and efforts needed to download and process the huge amounts of data locally.

The algorithm could reveal changing activity patterns of volcanoes at an early stage. This kind of model-based flagging gives clues about potentially important volcanic events, especially harbingers of impending eruptions. As a result, it enables decisions to be made in time on how to proceed. The automatic analysis and interpretation of the available satellite data can also help researchers to describe the periodic behavior of volcanoes in more detail in the long term.

Binayak Ghosh

Doctoral Candidate | Data Scientist | Geoscientist