Documentation#
Find your way around CryoStack.
CryoStack is the umbrella platform that connects cryosphere modeling and data-assimilation applications to reproducible software environments and computing resources. This page is the map — pick where to go next.
New to CryoStack?
CryoStack applications share one shape: configure, execute, monitor, visualize, download. Start with an application.
Getting Started
New here? Start with a short CryoStack orientation and find the right application for your work — open Getting Started.
CryoStack overview
What CryoStack is and how the browser, applications, environments, and backends fit together — see the home page and About.
CryoLauncher Getting Started
Run your first ice-sheet simulation and inspect the structured results — open the guide.
User Manual
The operational reference for configuring, running, and retrieving results across applications and execution environments — open the manual.
Choose an application.
Each application maintains its own documentation. Maturity differs — see the badges.
CryoLauncher
Run supported ice-sheet models through a browser interface, with a guided or an advanced workspace workflow, on remote and HPC resources.
ICESEE
Run ensemble data-assimilation workflows with supported numerical models for state estimation and parameter inference.
LIVIST
Explore Antarctic ice-sheet temperature products inferred from radar observations and constrained by borehole measurements.
Frozen Legacies
Explore historical Antarctic airborne radar surveys, flight tracks, and processed SPRI–NSF–TUD campaign observations.
Where runs execute.
Applications connect to computing resources through consistent execution backends. Details and settings are in the CryoLauncher User Manual.
Remote / HPC Supported
Linux servers and Slurm-managed clusters (such as Georgia Tech PACE) over SSH or the CryoStack Connector — execution guide.
Containers Supported
Digest-pinned tested Docker / OCI images and local SIF builds — see ICESEE-Containers.
ICESEE-Spack
Managed Spack environments with a first-time check-and-prepare workflow — preparing runs, ICESEE-Spack.
Cloud Supported
Bring-your-own AWS account, connected once through a CloudFormation role and used with temporary credentials only, on AWS Batch with a default Fargate compute mode and an Advanced EC2 alternative. CryoLauncher's ISSM and Icepack workflows have each completed end-to-end execution on both Fargate and EC2 On-Demand — for ISSM, MPI-parallel solver execution, institutional MATLAB licensing through Connector/Relay, postprocessing, result retrieval, and visualization; ICESEE's Lorenz-96 example has done the same on Fargate. EC2 Spot/GPU/multi-node and ICESEE's own Cloud/ISSM workflows remain unvalidated — see the Cloud Run Guide.
How you work in CryoLauncher.
Basic & Advanced
Guided, validated configuration (Basic) versus a user-owned workspace editor (Advanced).
My Workspace
Private, persistent, user-isolated examples and files; application examples stay read-only — examples.
Datasets
A reusable dataset area, referenced from examples and staged with the run — dataset guide.
Results & visualization
Structured result discovery and deterministic Solution / Field / Timestep rendering — results, visualization.
Building on CryoStack.
CryoStack separates the web interface, application layer, execution backends, and scientific data sources so each can evolve independently.
Source code
The gateway, application layer, connector, and deployment live in the CryoStack repository, under the ICESEE project.
Look something up.
Resources — an ecosystem-wide index of
applications, models, environments, datasets, and repositories.
Cloud Run Guide — connecting your own
AWS account and launching, monitoring, and retrieving a cloud run.
Troubleshooting —
common CryoLauncher issues and fixes.
Citation and
About — why CryoStack exists, design principles,
and how to cite it.