Technology
We're developing new approaches to forest carbon baseline assessment, combining geospatial AI with rigorous causal inference methods.
Our Principles
Transparency
We publish our methods, share our data, and welcome scrutiny. Science thrives on openness.
Science
Our work follows peer-reviewed standards and best practices in ecological research.
Independence
We provide objective assessments free from conflicts of interest.
Impact
Our science directly informs conservation action and policy decisions.
Challenging Current Methods
Our team's Science eLetter critiquing the synthetic control methods used to evaluate forest carbon projects—highlighting issues with donor pool selection and data accuracy that undermine conclusions about project effectiveness.
Read the full eLetter→
Map showing donor pool locations for a Colombian forest carbon project—highest-weighted donors were 900km away in the Amazon, in different ecoregions.
Our Technical Approach to Forest Monitoring
We're bringing together advances in machine learning and established econometric methods to tackle the counterfactual problem in forest carbon.
Geospatial Foundation Models
Leveraging pre-trained models on Earth observation data to extract rich representations of landscape characteristics, forest condition, and deforestation risk.
Causal Inference Methods
Applying techniques like synthetic control, matching, and difference-in-differences to construct credible counterfactual scenarios for carbon project assessment.
Carbon Stock Estimation
Combining satellite-derived biomass products with ground-truth data to estimate forest carbon stocks and their changes over time.
Scalable Analysis
Building reproducible workflows that can assess projects across diverse geographies and forest types, from tropical to temperate systems.
Want to learn more about our work?
We're always interested in connecting with others working on forest carbon science and counterfactual analysis.
Get in Touch