EO-driven MRV for Carbon Farming
Linking crop yield prediction to SOC change
Making the transition to regenerative agriculture measurable and bankable.
The project
The service aims to overcome the current economic paradox of the transition to regenerative agriculture: adopting new practices involves significant up-front costs (new agronomic methods, possible temporary yield reductions, training and production reorganization) against benefits that materialise only over the medium-to-long term.
From a banking perspective, this timeline reads as increased risk, restricting access to credit precisely when capital is most needed. Closing this gap requires scalable Monitoring, Reporting and Verification (MRV) approaches: methods able to measure environmental benefits in a consistent, reproducible way and translate them into credible economic and financial signals.
It is also a regulatory imperative, as the EU Carbon Removal and Carbon Farming framework (Regulation (EU) 2024/3012) sets out clear principles of quantification, monitoring, additionality and conservativeness. The work is carried out by MEEO together with Green Finance 4 Earth.
MEEO’s role
MEEO's contribution is a hybrid modelling pipeline that combines Machine Learning (ML) with a RothC-class process-based model for Soil Organic Carbon.
- Input: heterogeneous sources are integrated: Sentinel-2 satellite time series, soil maps (grids), climate projections, NDVI historical series and yield historical series. An AI engine (Random Forest combined with a cover-crop algorithm) estimates yield and biomass; an agronomic translator converts these into carbon inputs and root allocation, which feed the Rothamsted Carbon Model (RothC) simulation.
- Scenarios: the model estimates how SOC will evolve over a 10-year horizon under two scenarios: a baseline scenario (business as usual, conventional practices) and a regenerative scenario (cover crops). The difference in SOC between the two scenarios is translated into the amount of CO₂ sequestered, and then into a quantitative indicator of the farm's economic resilience.
- Financial translation: by modelling forward-looking scenarios that incorporate climate exposure, yield changes and carbon-related revenues, the framework estimates their impact on Probability of Default (PD) and Loss Given Default (LGD), and links biophysical outcomes to transition costs, land-value effects, break-even time and Net Present Value (NPV). In this way, environmental benefits are turned into bank-aligned indicators that support additionality assessment, investment appraisal and credit-risk evaluation.
The goal is to give farmers, financial institutions and policy makers a transparent, reproducible decision-support tool that makes the regenerative transition bankable.
Roadmap
The work carried out so far has demonstrated that climate and EO data, processed within the model, allow a consistent and reproducible estimation of SOC evolution over a 10-year horizon.
Future developments include the calibration and validation of the model through targeted in-situ sampling, guided by the areas of greatest uncertainty identified by the model itself, in order to improve its accuracy and predictive robustness.