How does Kanop help track land use changes in my sourcing regions?
Kanop provides continuous monitoring of land use changes across your sourcing regions using satellite imagery and AI analytics. Our platform tracks deforestation, forest degradation, and land conversion in real-time, enabling you to identify risks and compliance issues before they impact your operations.
We establish dynamic baselines for each sourcing location, allowing you to measure changes over time and quantify the environmental impact of your supply chain activities. This approach helps you demonstrate due diligence for regulations like EUDR, track progress toward sustainability commitments, and make informed sourcing decisions based on accurate, up-to-date environmental data.
What project type do you support?
Kanop supports ARR (tree planting/restoration), REDD+ (forest conservation), Blue Carbon (mangroves/wetlands), Improved Forest Management, Agroforestry, and Grassland restoration projects. We also provide supply chain monitoring and land use tracking.
Is Kanop compatible with Verra, Gold Standard and other standards?
Yes. Kanop's platform supports all major carbon standards including Verra VCS (VM0047, VM0048, VM0033), Gold Standard, BioCarbon Standard, Cercarbono, Isometric, and Equitable Earth.
We were recently selected by Isometric as a nature-based service partner across both feasibility assessment and earth-observation data, one of only four companies chosen in both categories. Projects certifying under Isometric's Biosphere protocols for reforestation, improved forest management, mangrove restoration, and agroforestry can integrate Kanop's monitoring directly within Isometric's certification workflow.
We are also an approved data provider for Equitable Earth, supplying satellite-derived earth-observation data directly into their certification pipeline. We support their M001 methodology for terrestrial forest restoration with automated eligibility assessments and dynamic baseline simulations.
Our satellite-based MRV technology aligns with methodological requirements throughout the entire project lifecycle, from eligibility assessment through baseline establishment to ongoing monitoring and verification. All measurements are scientifically validated and fully auditable, delivering the transparency and accuracy essential for successful project certification and investor confidence.
Can your solution differentiate between primary forests and planted forests?
Yes, our platform harnesses satellite imagery and AI to discern various types of forests, be it primary forests or plantations.
Is ground data necessary?
While our AI models deliver accurate results without ground data, incorporating your field measurements can significantly enhance precision and reduce uncertainty. This recalibration process is especially valuable for meeting specific certification requirements such as those in Verra's VM0047 methodology. Our platform makes integrating field data straightforward, allowing you to maximize both accuracy and cost-efficiency based on your project's specific needs.
What data sources do your data products use?
Kanop's models combine multiple satellite data sources to produce accurate biomass, canopy height, and forest cover estimates.
Optical imagery: Sentinel-2 (10-20m resolution, from 2015) and Landsat-8/9 (30m resolution, from 2013) provide spectral reflectance data linked to vegetation density and structure.
SAR (Synthetic Aperture Radar): Sentinel-1 (C-band) and PALSAR-2 (L-band) capture structural information about vegetation regardless of cloud cover or lighting conditions, making them essential for consistent monitoring in tropical regions.
Elevation data: The SRTM digital elevation model provides topographic context (slope, aspect, elevation) that informs biomass predictions.
Training data: Our deep learning models are trained on 125 million hectares of LiDAR reference data, including 42 million hectares from 754 airborne LiDAR campaigns and 83 million hectares from NASA's GEDI spaceborne LiDAR mission, spanning 76 countries.
Field measurements (optional): When available, user-provided plot-level measurements can be used to recalibrate model outputs through a Gaussian Process Regression approach, improving local accuracy.
How accurate is the estimation of the aboveground biomass?
Accuracy depends on the scale of analysis. In carbon accounting and MRV, the relevant metric is typically the average biomass over a project area or polygon, not individual pixel values.
At the site level (the scale most relevant to carbon accounting), our model achieves an RMSE of 41.7 t DM/ha (relative RMSE = 27%) and an R² of 0.73, validated across 110 independent sites covering over 1,000,000 hectares of LiDAR-derived reference data from peer-reviewed publications. This means the model reliably estimates the average carbon stock of a project area.
At the pixel level (30m resolution), the global weighted RMSE is 94.9 t DM/ha. This is roughly double the site-level figure, which is expected: fine-scale errors from canopy gaps, mixed pixels, and terrain effects partially cancel out when averaged over larger areas.
Every estimate includes a confidence interval (95% by default), propagated from pixel to polygon level and accounting for spatial correlation between pixels. When field measurements are available, an optional post-calibration step using Gaussian Process Regression can further reduce uncertainty at the local level.
