Caldera volcanoes—the ones that produce Earth’s most explosive eruptions—can deform for decades before they blow, but scientists cannot yet tell whether that swelling signals an imminent eruption or just background rumblings. This project aims to crack that ambiguity by building a unified model that links magma movement, crustal deformation, and underground hydrothermal systems. Current monitoring relies heavily on satellite and ground-based geodesy—measuring how the ground bulges or sinks—but the models used to interpret those signals are too simple. They ignore how magma ascends through a crystal mush, how it stalls, and how hot fluids in the shallow crust distort the surface. Without that complexity, observatories cannot confidently issue evacuation orders. The team will combine physical experiments, petrological data from past eruptions, and Bayesian inversion software to create a holistic model. They will test it on the Diamante-Maipo caldera in Chile and Argentina, then extend it to other calderas worldwide. If successful, the work will give volcano observatories in Chile, Argentina, and the US a practical tool to interpret unrest signals in real time, improving eruption forecasts and reducing false alarms that erode public trust.
View original technical description
This project will deliver a step-change in understanding the origin of surface deformation at complex and dangerous caldera volcanoes. Highly explosive caldera-forming eruptions are infrequent but responsible for some of the most catastrophic geological events, with potentially devastating local, regional and global consequences. By unifying the latest insights from our state-of-the-art analogue and numerical models with petrological datasets and geodetic inversion algorithms we will improve the ability to predict eruptions at caldera volcanoes. We need to advance the capabilities of the inverse modelling approach by linking analogue models and geodetic volcano monitoring methods and integrating complex crust and magma rheologies. Geodesy is a major tool used by volcano observatories to assess volcanic unrest, and known model discrepancies must be improved to inform eruption assessments. It is the recognition of magma ascent at a caldera which can trigger evacuation orders, but to improve eruption assessments we need to understand how magma ascent in a crystal mush evolves at caldera systems and how these affect surface displacements in the presence of a shallow crustal hydrothermal system. By combining our physical and chemical models with the new insights on the different timescales of magma ascent and stalling across the lifetime of an active caldera, we will enhance and improve interpretations of caldera unrest signals. Our holistic physical, chemical and hydrothermal models will be used to inform volcano monitoring network deployment campaigns in Chile (SERNAGEOMIN), Argentina (CONICET-SEGEMAR) and across the Americas (USGS), with education, communication, EDI and ethical practices embedded across our research programme. We will apply simultaneous surface and sub-surface imaging techniques on caldera analogue experiments and use the Geodetic Bayesian Inversion Software (GBIS) to place new constraints on magma intrusion and host-rock deformation source parameters that can be applied to natural datasets. We will collect compositional data from the crystal cargo of stratigraphically-constrained eruptions from the Diamante-Maipo system and apply thermobarometry and a range of geochronometric approaches to understand the locations and timescales of magma accumulation and ascent. We will create a numerical simulation of magma intrusion within a magma mush to resolve its impact on surface distortions in the presence of a shallow crustal hydrothermal system at a model caldera volcano. We will apply our holistic physical-chemical-thermal model to Diamante-Maipo and test its capabilities for scenario mapping for other caldera volcanoes around the world. Building on our track record in pioneering magma ascent modelling and leading multidisciplinary projects with academia-observatory collaborations, our models will have rapid impact by informing on the deployment of volcano monitoring equipment.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know