Skip to content

Separating Magmatic and Hydrothermal Sources in Volcanic Deformation

Edna W. Dualeh 1, Juliet Biggs1

Abstract

Determining whether unrest at silicic systems is primarily driven by magmatic, hydrothermal, or a combination of both is crucial for assessing eruptive hazards, although remains a long-standing debate at many well-studied systems. Recent advances in machine learning and satellite technology, especially through the Sentinel-1 mission, allow for more detailed analysis of spatial and temporal deformation patterns, enabling us to separate signals with different characteristics. In this study, we use spatial Independent Component Analysis (sICA) to separate the magmatic and hydrothermal contribution to surface deformation based on their different spatial characteristics. Using Corbetti Caldera, Ethiopia, as a case study, we extract two distinct surface deformation patterns, related to a steady uplift of 5.1 cm/year linked to a shallow magmatic source, and a laterally bound hydrothermal signal with a seasonal amplitude of 0.65 cm. We found that while the original data can be modelled well using single Mogi point source, sICA demonstrated that a two-source model provides a better solution. These findings highlight the potential of combining high-spatial and temporal resolution datasets with machine learning techniques to separate magmatic and hydrothermal processes in the surface deformation. Understanding the cause of unrest provides critical insights for hazard assessment.