Conservation Area Prioritization
Through Artificial INtelligence

The Captain Project

How do we best protect biodiversity in a rapidly changing world and with limited resources?

Over a million species face extinction, carrying with them untold options for food medicine, fibre, shelter, ecological resilience, aesthetic and cultural values. We urgently need to design conservation policies that maximize the protection of biodiversity and its contributions to people, within the constraints of limited budgets.


Harnessing the power of AI to optimize conservation efforts

We use reinforcement learning to train models for conservation prioritization that best use the available data and resources. CAPTAIN models can work with basic species distribution data but can handle complex multidimensional data and their temporal trends, including land use and climate change.

Captain flow

Conservation policies outperforming the state-of-the-art

Our experiments using simulated and empirical data indicate that CAPTAIN yields more reliable conservation solutions than alternative state-of-the-art software for systematic conservation planning.

Captain vs Marxan performance

Customized prioritization targets

Optimize policies toward different conservation targets, e.g. aiming to minimize species loss or to maximize the amount of protected area, and compare their outcomes and tradeoffs.

Minimize species loss
Minimize species loss
Maximize carbon storage
Minimize economic value loss
Maximize protected area
Maximize protected area


A desktop app is coming soon. The Python source code is available on GitHub.

We provide Jupyter Notebooks that showcase the capabilities of Captain.

Join the community

Ask questions and get help on GitHub Discussions.