Data-Driven Estimation of Carbon Sequestration Value for Climate Change Mitigation and Land Restoration in Sub-Saharan Africa

Land restoration and reforestation are widely recognized as cost-effective climate mitigation strategies, yet across Sub-Saharan Africa, land with genuine carbon sequestration potential frequently goes unmonetized because verifying and quantifying that potential has traditionally required expensive field surveys and third-party assessments. This creates a persistent gap: degraded or restorable land is rarely brought into carbon markets or attracts climate finance, not because the ecological value is not there, but because it cannot be measured and verified cheaply or credibly at scale.
This project addresses that gap directly by building a machine learning pipeline that estimates land-level carbon sequestration potential using entirely free, open-access satellite and environmental datasets. Using Sentinel-2 and Landsat imagery to derive vegetation and biomass proxies (via NDVI and related indices), MODIS vegetation index time series to track land cover change and restoration progress, SoilGrids data for soil organic carbon content, and CHIRPS rainfall data to contextualize vegetation growth potential, the project trains models to predict carbon sequestration capacity across selected pilot regions in Sub-Saharan Africa.

Model outputs are then translated into economic terms indicative carbon credit value ranges using publicly available carbon market reference pricing and methodology documentation (e.g., Verra and Gold Standard project frameworks), positioning the work at the intersection of environmental science, data science, and environmental economics. Rather than treating carbon valuation as a one-off assessment, the project produces a reusable, low-cost pipeline that can be applied to new land parcels or regions using only open data, substantially lowering the barrier to bringing restoration projects into climate finance mechanisms.

The project is expected to produce a peer-reviewed manuscript, an open-source reproducible codebase, and a policy brief aimed at climate finance and land-restoration stakeholders, demonstrating a scalable, data-driven pathway for converting environmental restoration into economic value.

Open to graduate students, early-career researchers, and academics with backgrounds in data science, remote sensing, climate science, or environmental economics. No institutional restriction.

Active, consistent participation with regular check-ins (e.g., biweekly); genuine contribution to data work, modelling, or writing; authorship credit based strictly on demonstrated contribution; respectful, collaborative communication throughout.