OpenForest

Branching Out: New OpenForest Catalog Advances AI for Ecosystem Monitoring

Dynamic database compiles global forest datasets to enhance machine learning applications.

Forests worldwide face growing threats from human activities, and monitoring the impact at scale remains a significant challenge. 

A new research paper, “OpenForest: A Data Catalog for Machine Learning in Forest Monitoring,” introduces an open-access repository aimed at improving the accessibility and usability of forest data for artificial intelligence (AI) applications.

Led by Arthur Ouaknine of McGill University and Mila – Quebec AI Institute, along with co-authors Teja Kattenborn, Etienne Laliberté, and David Rolnick, the study compiles 86 open-access datasets spanning satellite imagery, aerial surveys, and ground-based observations. OpenForest serves as a centralized hub for forest data, fostering interdisciplinary research and driving innovation in AI-based forest monitoring.

The launch of OpenForest aligns with the mission of the AI and Biodiversity Change Global Center, which seeks to leverage AI for large-scale environmental challenges. AI is increasingly applied to tree species identification, biomass estimation, and deforestation tracking, yet its progress depends on diverse, high-quality data. OpenForest removes a major hurdle by aggregating scattered datasets into a single, open-source platform, enabling researchers to train machine learning models with more robust and representative data.

“This is a rapidly growing field at the intersection of AI, land management, and biodiversity conservation,” said David Rolnick, co-author and Canadian lead of the AI and Biodiversity Change Global Center. “By making forest data more accessible, OpenForest empowers researchers to develop better AI-driven monitoring solutions.”

The OpenForest catalog encompasses spatially and temporally diverse datasets, covering everything from individual tree inventories to continent-scale satellite observations. By integrating these resources, the platform supports improved AI applications for forest classification, carbon stock assessment, and conservation efforts.

Tanya Berger-Wolf, ABC Global Center American lead, praised the initiative. 

“This paper showcases a multimodal open forest dataset compiled by the AI and Biodiversity Change Global Center. Open data is a game-changer for biodiversity science," she said.

The OpenForest team encourages contributions from researchers and institutions to expand the dataset collection. Future iterations aim to incorporate real-time data streams, enhanced metadata, and AI-assisted curation tools.

“This is just the beginning,” Ouaknine said. “We hope OpenForest will become an essential tool for AI researchers, ecologists, and policymakers working toward sustainable forest management.”

OpenForest is available at GitHub, and the full research paper is published in Environmental Data Science.