Elmer/Ice

Elmer/Ice
Name	Elmer/Ice
Developer	ElmerSolver consortium, CSC – IT Center for Science, European Commission
Released	2000s
Programming language	C++, Python
Operating system	Linux, Windows, macOS
License	GNU General Public License

Elmer/Ice Elmer/Ice is a modular computational framework designed for simulation of glaciology, ice sheet dynamics, and coupled cryospheric processes. It integrates finite element solvers with physics modules to model flow, thermodynamics, and interaction with atmosphere and ocean forcings, supporting researchers and institutions working on glacial modeling and climate-related studies. The project is connected to a network of scientific institutions and leverages interoperable standards to interact with community tools and datasets.

Overview

Elmer/Ice originated from collaborations among the CSC – IT Center for Science, European Space Agency, University of Oslo, MET Norway, and regional cryosphere groups to provide a research-grade toolchain for ice dynamics. The framework builds on the Elmer multi-physics suite and extends it with domain-specific formulations for shallow ice, higher-order, and full-Stokes approximations, enabling coupling to forcings from ERA5, ECMWF, Copernicus datasets. Elmer/Ice supports meshes from preprocessing tools such as Gmsh, SALOME, and integrates with visualization systems like ParaView and Visit for post-processing.

Architecture

Elmer/Ice is structured as a set of modular components: a finite element kernel derived from Elmer, constitutive-law modules for non-Newtonian rheology, solvers for thermomechanical coupling, and interfaces for boundary conditions driven by observational products like MODIS, ICESat, GRACE. The architecture uses mesh representations compatible with Gmsh and NetCDF-based grids and adopts parallelization via MPI and OpenMP to run on cluster resources provided by PRACE and national supercomputing centers. Interoperability layers allow scripting with Python bindings and integration into workflows orchestrated by Snakemake or CMake-based build systems.

Features and Capabilities

The codebase implements multiple flow approximations: shallow-ice approximation (SIA), shallow-shelf approximation (SSA), higher-order models, and full-Stokes solvers equivalent to formulations used in major field campaigns. It includes thermodynamics modules for cold and temperate ice, subglacial hydrology parameterizations, and calving criteria informed by remote sensing products such as Sentinel-1, Landsat, and ICESat-2. Elmer/Ice supports data assimilation using adjoint-based techniques and ensemble approaches compatible with PEST and DART. For boundary forcings it ingests atmospheric reanalyses like ERA-Interim and ERA5, ocean interactions from HYCOM and NEMO, and mass-balance inputs from MAR and RACMO.

Use Cases and Applications

Researchers use Elmer/Ice for regional studies of Greenland ice sheet dynamics, outlet glacier evolution in Antarctica, and sensitivity experiments linked to IPCC assessments and CMIP-derived scenarios. It has been applied to investigate grounding line migration influenced by Pine Island Glacier retreats, tide–glacier interaction studies in fjords informed by OSNAP observations, and subglacial drainage responses observed during Meltwater pulses. Operational centers and research groups employ Elmer/Ice in coupled cryosphere–ocean modeling chains alongside tools like ROMS, MITgcm, and glaciological datasets from NSIDC and PANGAEA repositories.

Performance and Scalability

Elmer/Ice leverages high-performance computing paradigms to scale from single-node workstations to national supercomputers such as those in the PRACE and Nordic e-Infrastructure Collaboration ecosystems. Its parallel solvers use domain decomposition and multigrid preconditioners adapted from the PETSc and Hypre libraries, enabling large-scale full-Stokes simulations on meshes with millions of degrees of freedom. Performance tuning draws on profiling tools such as gprof, VTune, and HPCToolkit; scalability studies compare runtime and memory footprints against community models like ISSM and Elmer benchmarks for ice dynamics.

Development and Community

Development is coordinated through academic consortia, code repositories, and collaborative platforms used by institutes including University of Cambridge, University of Bristol, University of Bergen, and national research labs. Contributions follow open-source practices common in projects hosted on services similar to GitLab and GitHub with issue tracking, continuous integration, and unit tests. The user community engages via mailing lists, workshops at venues such as the EGU General Assembly and AGU Fall Meeting, and training schools sponsored by groups like SCAR and IACS to ensure reproducibility and interoperability with community standards.

Security and Privacy

As an open scientific code, Elmer/Ice emphasizes reproducibility and transparent provenance rather than confidentiality, managing data inputs and outputs in standardized formats like NetCDF and HDF5 that facilitate archiving in repositories such as Zenodo and PANGAEA. For deployments on shared HPC resources overseen by PRACE or national centers, operational security relies on institutional policies from providers including CSC – IT Center for Science and access control via LDAP or token-based systems used by cloud platforms from vendors such as Amazon Web Services or Microsoft Azure. Data sensitivity concerns primarily address embargoed observational datasets from missions like ICESat-2 and coordinated access via project agreements.

Category:Computational science software