AI-enabled digital MRV for scaling agroforestry and trees outside forests in India

Agroforestry and trees outside forests (AF-TOF) are increasingly recognized as vital for climate resilience, soil restoration, and rural livelihoods in South Asia. Yet, their large-scale adoption is constrained by the absence of reliable, affordable, and transparent measurement, reporting, and verification (MRV) systems. Traditional field inventories are costly and inconsistent, limiting the credibility of carbon and ecosystem service claims. Young saplings below the age of 3 years are missed by most satellite products, and many existing solutions are not affordable or openly available, thus preventing ecosystem adoption at a larger scale. To address these gaps, we developed an open-source digital MRV framework that leverages drones and artificial intelligence to automate tree detection and carbon quantification in smallholder landscapes. Using high-resolution drone orthomosaics, tree crown identification was conducted with deep learning models for object detection, while Gaussian Process Regression was applied for above-ground biomass estimation. Models were trained and validated on ~120,000 manually annotated crowns across agroforestry sites in Maharashtra, India. Our results demonstrate substantial improvements in both accuracy and cost-efficiency compared to conventional survey methods. By releasing our work under open-source licenses, the framework reduces barriers to adoption, enables independent validation, and encourages adaptation by practitioners, researchers, and policymakers across geographies.

Farmers for Forests (F4F) addresses a structural failure at the intersection of climate change, land degradation, and rural poverty. Across India, smallholder and tribal farmers are among the most vulnerable to climate-induced shocks such as droughts, floods, and erratic rainfall. At the same time, rising global demand for carbon credits and ecosystem restoration offers an opportunity to channel climate finance to rural communities. Yet smallholders remain systematically excluded from Payments for Ecosystem Services (PES) and carbon markets. This exclusion is not due to lack of impact, but due to the failure of conventional Monitoring, Reporting, and Verification (MRV) systems. Existing MRV approaches are manual, expensive and slow, making fragmented, small landholdings financially unviable for climate finance. Early-stage restoration is particularly hard to verify: young trees are invisible to most satellite products, while field sampling relies on sparse plots and heavy extrapolation, creating high uncertainty among registries and funders. India has nearly 30% of its land degraded, and over 120 million smallholder farmers face declining soil fertility and livelihood insecurity. Agroforestry and forest protection are proven solutions that improve farm incomes, restore biodiversity, and sequester carbon - but adoption remains limited because impact cannot be measured credibly or affordably during the critical first 3–5 years. F4F solves this MRV bottleneck using AI-driven, high-resolution monitoring. We provide farmers with subsidized saplings, drip irrigation, organic inputs, and free training to transition to agroforestry. Using drone imagery and machine learning models, we perform tree-level and plot-level monitoring at approximately one-tenth the cost of traditional high-resolution systems. Verified plots are then connected to global carbon markets. Farmers benefit through diversified agroforestry produce, premium carbon revenues, lower input costs, improved soil health, and greater resilience. By making verification affordable, transparent, and scalable, F4F unlocks climate finance for millions of smallholders while accelerating ecosystem restoration at scale.