Polar TEP

Polar TEP (Polar Thematic Exploitation Platform) is a cloud working environment, developed by Polar View for the European Space Agency, for accessing and analysing Earth observation data of the polar regions without downloading large datasets. Satellites are the only consistent, year-round source of data over the remote Arctic and Antarctic, but the growing volume of data makes local download and processing inefficient.

Workspaces connect to any data repository with an API or standard OGC endpoints, including Sentinel data through the Copernicus Data Space Ecosystem, RADARSAT Constellation Mission (RCM) data through EODMS, Copernicus Marine Service metocean data and national ice charts. Depending on configuration, they provide Jupyter Notebooks in Python, QGIS, a Conda store, headless execution and Argo Workflows on Kubernetes, an MLflow machine-learning environment with GPU virtual machines, S3 storage with STAC catalogues, and information dashboards. Users can run built-in processors or deploy their own in Docker containers, and share code and results with their team.

The platform is open source and infrastructure agnostic. The Geological Survey of Denmark and Greenland has used it since 2017 to process Sentinel-3 ice sheet albedo in near real time, which Polar View says cut processing time by a factor of ten; the Norwegian Geotechnical Institute used it for a 2016-2026 reanalysis of landfast sea ice over Svalbard (5.5 TB of results); and 42 teams in the ESA Phi-Lab AutoICE sea-ice challenge used its machine-learning resources. Published monthly plans start at EUR 500 (development environments from EUR 550), with a three-month free trial of the Jupyter environment, and ESA Network of Resources sponsorship is available for research, education and pre-commercial use.