Tracking Magma Migration in Near-Real Time using Radar Interferometry - the FUTUREVOLC Supersite Approach
Hooper, Andrew1; Spaans, Karsten1; Sigmundsson, Freysteinn2
1University of Leeds, UNITED KINGDOM; 2University of Iceland, ICELAND

Our ability to predict the onset and evolution of eruptions depends, in part, on our ability to image the movement of magma beneath the surface. To date, this has been achieved primarily using GPS and seismic data gathered in situ, while synthetic aperture radar (SAR) data acquired from space have been used, in the main, to analyse eruptions retrospectively. The short revisit times of some current SAR missions mean that it is now plausible to use radar imaging also for the monitoring of volcanoes, while the upcoming Sentinel-1 mission will allow for routine monitoring of volcanoes globally. One of the aims of FUTUREVOLC, a collaborative project encompassing 26 partners in 10 countries, is to develop a system that will automatically ingest SAR images and output deformation maps in near-real time. FUTUREVOLC is a supersite project, where Iceland has been selected as the target area. In the past twenty years, SAR data from ERS and Envisat satellites have been used to constrain numerous deformation signals in Iceland associated with magma movement (and many other deformation processes, such as tectonic movement, the response to snow and ice load changes, geothermal exploitation and land sliding). Iceland thus provides an excellent natural laboratory for development of this system, but ultimately it should be applicable to all subaerially exposed volcanoes worldwide.

In order to achieve the goal of analysing SAR images in near-real time, a new approach is needed. We have developed a system that uses pre-analysis of the SAR archive to identify pixels that have similar noise characteristics, but not necessarily the same deformation history. This information can be used to very quickly identify coherent pixels in interferograms formed using new images acquired in a time of crisis. We have also developed new methods for 'unwrapping' the phase of these newly formed interferograms, which utilise the redundancy of the interferogram network to automatically and robustly detect unwrapping errors. We use data acquired prior to, and during, the 2010 Eyjafjallajoekull eruptions to test our algorithms. A subsequent aim of FUTUREVOLC will be to take the output from this SAR processing chain, together with in situ data, and produce models of magma migration and stress evolution, also in near-real time.