Forecasting glacier outburst floods with LUMI

The European High-Performance computing Centre of Excellence for Exascale on Solid Earth (European HPC Centre of Excellence ChEESE) is evaluating and forecasting geohazards, with civil protection as a primary objective. The ChEESE CoE addresses multiple challenges posed by geohazards and develops accurate predictions with the help of supercomputers including LUMI.

Rapidly-rising glacier outburst floods (jökulhlaups) present a challenging problem for modelling. Outbursts from subglacial geothermal areas, marginal lakes, and subglacial volcanic eruptions are common in Iceland. They pose a substantial hazard to settled areas as well as to communication lines, roads, and other infrastructure near glaciers.

CSC – IT Center for Science in Finland is one of the partners of the CoE. CSC is bringing its expertise to the CoE by contributing with the Elmer/Ice software. Together with the Icelandic Meteorological Office, CSC is implementing a model for simulations of jökulhlaups in Elmer/Ice and helping to port natural hazard applications to the LUMI supercomputer.

– This Centre of Excellence is a unique opportunity to promote the further development and porting of applications – CSC’s Elmer/Ice amongst them – related to natural hazards to Europe flagship computer, LUMI, and enable critical urgent computing in case of the event of a natural disaster in order to increase the safety of those people being affected, both inside Europe and all over the world, said Thomas Zwinger, Modelling Specialist at CSC.

Images: Glacier outburst floods, or jökulhlaups, from subglacial geothermal areas, marginal lakes and subglacial volcanic eruptions are common in Iceland and they pose a substantial hazard to settled areas as well as to roads, communication lines and other infrastructure near glaciers. This photo shows the ~2.5-km diameter Eastern Skaftá Cauldron after the ice on top of a subglacial lake subsided ~100 m during a jökulhlaup in September 2021. The accompanying map and discharge graph show the flood path on the southwestern side of the Vatnajökull ice cap and the variation of flood discharge with time for the last three floods in 2015, 2018 and 2021. Image copyright: Icelandic Met Office

The CoE has the following main objectives:

• To prepare 11 community flagship codes to address 12 domain-specific exascale Computational Challenges, enlarging the areas covered during the first implementation phase of the project with additional disciplines.
• To develop a new generation of 9 Pilot Demonstrators for scientific problems requiring exascale computing.
• To use these Pilot Demonstrators in 15 Simulation Cases of particular relevance in terms of science, social relevance, or urgency.

– We are glad about this opportunity of cooperation with a large number of partners from academics and industry to utilize the investments in European HPC infrastructure in connection to natural hazard forecasting, said Jussi Heikonen, Development Manager at CSC.

– The ChEESE iniative is enhancing the capabilities of European Solid Earth community and enables us to provide services related to urgent computing, early warning, geohazard assessment, and data analysis for European civil protections, said Arnau Folch, CSIC Professor and ChEESE project coordinator.

The CoE ChEESE is coordinated by the Spanish National Research Council (CSIC), being co-funded with more than €7 million by the European Commission and national funds from ten other European countries. The new CoE ChEESE is the second phase of a successful research project that was coordinated by the Barcelona Supercomputing Center until 2022.

A range of world-class academic and industry partners from across Europe are participating in the CoE, such as: Atos, Barcelona Supercomputing Center, Centre National de la Recherche Scientifique, CINECA, CSC – IT Center for Science, Instituto Nationale di Geofísica e Vulcanologia, Institut de Physique du Globe de Paris, Johannes Gutenberg-Universitat Mainz, Lundwig-Maximilians-Universität Münich, Ruder Boskovic Institute, Sorbonne Universite, Stiftelsen Norges Geotekniske Institutt, Technical University of Munich, Universidad de Málaga, Unversität Stuttgart and the Icelandic Meteorological Office.

This article was originally published on the CSC website.

Read also:

Interview: Thomas Zwinger explains how the ChEESE Project is enhacing glacier modeling and risk prediction