Some volcanoes can switch from gentle lava flows to violent explosions within minutes, yet no existing model predicts these dangerous transitions. This project targets that gap by studying three Caribbean volcanoes—La Soufrière in St Vincent, Soufrière Hills in Montserrat, and Mt Pelée in Martinique—that have a history of unsteady, deadly eruptions. The team will combine detailed analysis of eruption deposits, seismic data from ground sensors, and machine learning to build the first computer models that capture how conditions change inside a volcano in real time. If successful, these models could give monitoring agencies in the Eastern Caribbean—and volcanic regions worldwide—a practical tool to forecast when a quiet eruption is about to escalate. That would allow earlier warnings, better evacuation decisions, and fewer lives lost. The project also trains a new generation of researchers and strengthens partnerships with Caribbean hazard managers, ensuring the science translates into real monitoring improvements. While the work is fundamentally about understanding volcanic physics, its direct outcome is a predictive capability that currently does not exist.
View original technical description
A grand challenge for volcanology is to understand the drivers of eruptive transitions. Many volcanoes transition between gently effusive and violently explosive behaviour, or show rapid fluctuations in explosivity during eruption, leading to dangerous escalations in eruptive hazard. Unsteady eruptions are common, but are currently hard to anticipate as there are no models that capture this behaviour. Because of this, unsteady eruptions are associated with significant disruption and fatalities in both historical and contemporary records. Our multidisciplinary team provides the expertise required to integrate and analyse diverse datasets and create models for unsteady volcanic systems that will realise new capabilities for predicting dangerous volcanic eruptions. Our aim is to deliver a step-change in how we understand and anticipate the dynamic drivers for ‘dangerous’ eruptive transitions. We can do this now because of recent analytical, theoretical and instrumental developments in volcanology, much of it led by Ex-X team members, and acceleration in numerical modelling capacities. Our collaborations with key international partners who have generated enabling data and models are fundamental to this innovation. Together, we will forensically examine eruptions from three volcanoes with rich records of unsteady and dangerous eruptions (La Soufrière, St Vincent; Soufrière Hills Volcano, Montserrat; Mt Pelée, Martinique). Their deep storage systems and magmatic characters are well-understood, so we can target critical knowledge gaps arising from their unsteady, time-dependent behaviour. These volcanoes provide archetypes for dangerous eruptive transitions worldwide, and our insights will directly benefit interpretations of ongoing unrest and future activity in the Caribbean, and beyond. The Large Grant format allows us to bring together the diverse expertise needed to do this well, and in an innovative and timely fashion. Our project has the following objectives: (1) to develop new methodologies that capture fluctuating conduit input (three phase conduit flow, loading and erosion) and consequent variations in eruptive behaviour by: (a) using ‘microstratigraphies’ to capture the spatial and temporal changes that drive transitions which are uniquely recorded in the eruptive deposits and (b) using nodal seismometers and machine learning to enhance the spatial and temporal resolution of seismic and geophysical records of eruptions and their changing conditions (2) to develop models with the essential time-dependence needed to describe unsteadiness in each part of the system, beginning with the current state-of-the-art knowledge for: (a) disequilibrium conduit flow and (b) unsteady eruption columns (3) to create an end-to-end description of the drivers of past eruptive transitions through a brand-new coupled model capable of capturing fragmentation and column collapse, validated and refined via our physically derived datasets. (4) to demonstrate how this knowledge can improve monitoring and warning systems in the Eastern Caribbean, and beyond by (a) using the coupled model to predict geophysical precursors to transitions (b) using the combined datasets to evaluate the range of likely eruptive scenarios and trajectories that may lead to dangerous eruptive transitions at Eastern Caribbean volcanoes. ? (5) Through our engagement with Eastern Caribbean partners, and our attention to the career development of early career researchers, we will support and develop a new generation of researchers and partnerships capable of tackling important multidisciplinary problems in volcanology, volcano monitoring and management. Ex-X will particularly benefit those responsible for volcanic hazard monitoring and management, and through them the exposed populations and managers of risk in volcanic countries.
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