
The Global Ecosystems Atlas is convened by the Group on Earth Observations (GEO), which together with the Global Ecology Lab at James Cook University forms the Atlas core team. In addition to the core team is the Technical Consortium that brings together IUCN, Ai2, ESA, Esri, Google to deliver end‑to‑end capability for ecosystem mapping: from training data and modelling to integration and global access.
The core team and the Consortium have provided technical expertise and resources that have directly informed the design and implementation of the Atlas systems, methods, and resources, ensuring scientific and technical robustness.
AI2 developed OlmoEarth, a browser‑based workbench for rapid, high‑volume contributions of labelled ecosystem points, AI‑enhanced feedback, and active‑learning. This supports the development of EcoTrain, the structured, open training data aligned with the IUCN Global Ecosystem Typology. OlmoEarth also delivers state-of-the-art foundation models trained on multi‑sensor satellite data to identify patterns in the landscape and generate high-quality, high-resolution ecosystem maps using known occurrences of ecosystems.
ESA contributes through the World Ecosystem Extent Dynamics (WEED) project. Through the WEED project, ESA contributes cross-referencing outputs and national‑level ecosystem maps following IUCN Global Ecosystems Typology. In addition, ESA plays a coordination function in Europe, supporting direct engagement with European countries, through the European Commission, the European Environment Agency (EEA) and the CBD Technical and Scientific Regional Cooperation Support Centre (CBD-TSC) for Europe.
Esri provides Survey123, the structured field data collection tool used by the Atlas field team. Esri also developed the custom geoprocessing pipelines to synthesise existing spatial data to produce the Atlas Data Product Suite. This includes scripted workflows for cloud-based preprocessing, re‑projecting, reclassifying and mosaicking inputs to produce the harmonised Atlas synthesis map and accompanying statistical summaries. Once complete, the synthesis map will be disseminated through the ArcGIS Living Atlas to users, governments, investors and conservationists worldwide.
Google identified over 90,000 stratified random sample points to support the creation of the EcoTrain dataset. The Global Ecosystems Atlas modelling stream relies on AlphaEarth Foundations, the state-of-the-art geospatial foundation model built with multimodal satellite and environmental data to accurately and efficiently describe broad spatial patterns. Models are implemented in Google Earth Engine, the cloud-based geoprocessing platform. Once complete, the synthesis map will be disseminated as an asset through the Google Earth Engine Data Catalogue, to users worldwide.
IUCN ensures consistency with the IUCN Global Ecosystem Typology through technical standards and cross-referencing protocols. IUCN supports international coordination between IUCN Regional offices, the implementers of Red List of Ecosystems (RLE) assessments, the Commission on Ecosystem Management (CEM) and its Red List of Ecosystems thematic group, including supporting the submission of RLE datasets to the Global Ecosystems Atlas. The IUCN also supports capacity building through training workshops, outreach, and technical engagement via IUCN Academy.