Skip to content

Data Assimilation for Probabilistic Forecasting of Lava Flows in Real-Time

Louise Cordrie 1, Antonio Costa 1, Giovanni Macedonio 2, Francesco Zuccarello 3, Gaetana Ganci 3, Annalisa Cappello 3, Roberto Spina2

  • Affiliations: 1 Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna, Bologna, Italy; 2 Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano, Napoli, Italy; 3 Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Catania, Catania, Italy;

  • Presentation type: Poster

  • Presentation time: Tuesday 16:30 - 18:30, Room Poster Hall

  • Poster Board Number: 43

  • Programme No: 6.5.12

  • Theme 6 > Session 5

Abstract

The monitoring of effusive volcanic eruptions involves processing multiple signals (volcanic tremors, ground deformation, satellite imagery...). However, the uncertainties inherent in these data, combined with the complexity of lava flow physics, make precise deterministic forecasts of lava flow propagation highly challenging. An ensemble-based probabilistic approach, combined with a robust numerical model of lava flow, addresses the uncertainties related to lava rheology and propagation. By integrating data assimilation techniques, this approach also accounts for uncertainties in observational data, producing data-informed forecasts. Using these methods, a real-time workflow has been developed to forecast lava flow propagation during eruptions by continuously incorporating new incoming data. Built around VLAVA, a new numerical model that simulates temperature-dependent viscous lava flow propagation over a complex topography, the workflow generates multiple combinations of model source parameters, creating an ensemble of lava flow simulations. The weight associated with each member of the ensemble can be adjusted in real-time based on their similarity to the observations. The aggregation of these weighted lava flow simulations results in updated probabilistic forecasts. In addition, lava fields extracted from satellite data can be directly assimilated into simulations, driving the model toward even more realistic lava flow predictions. This data-assimilation method has been coupled with satellite images processing and uncertainty evaluation modules which are key points for the workflow's accuracy and robustness. These developments have shown promising results when applied to historical eruptions of Mount Etna (Italy) and will be extended to other volcanoes, including Fagradalsfjall (Iceland).