Introduction

Forest carbon projects are critical tools for addressing climate change, attracting climate finance, advancing conservation initiatives, and supporting environmental justice. Their success, however, depends heavily on the credibility of the carbon measurement, monitoring, reporting, and verification (MMRV) strategies they use. Namely, how the project proposes to conservatively measure carbon stocks, monitor them through time, and report their estimates transparently. 

MMRV is fundamental to credit quality. If the methods aren't accurate, transparent, and defensible, a project’s carbon stock estimates won't be considered high-quality and won't garner investor interest. 

Traditional forest carbon estimation is hard to scale

Manually measuring trees has been a cornerstone of forest management and research for over 100 years, helping to build an understanding of forestry fundamentals like tree growth, mortality, and recruitment through time. 

Field crews collect data on tree species, height, and diameter to estimate above-ground biomass and carbon stocks using allometric equations—models that relate measurable tree attributes to harder-to-measure attributes like above-ground biomass and carbon.

Two key challenges exist in this traditional process. 

• Manually measuring trees is expensive, time-consuming, and logistically challenging, especially for large and remote projects. 

• Allometric models don’t exist for every species or forest type, creating uncertainty around when, where, and how they should be applied. 

Remote sensing enables scalable forest carbon estimation

To overcome the limitations of manual measurement, project developers and credit registries are increasingly integrating remote sensing datasets and AI tools (e.g., satellite imagery and predictive models) to supplement their field measurement approaches. Using remote sensing in combination with field measurements helps to:

• Improve the cost-effectiveness of carbon stock estimation over space and time.

• Promote the quality and defensibility of crediting frameworks.

• Increase purchaser and investor confidence in project performance.

Global satellite datasets like Landsat cover the entire planet and have provided consistent Earth observations for the past 40 years. The potential benefits of remote sensing and AI are numerous, but it's important to also recognize that they are not a silver bullet for ecosystem assessment. Ground-based data will always be needed, and local experience and expertise should remain a central element to forest mapping and monitoring. Combining these data types and experience is critical to effective and equitable management.

A hallmark of remote sensing data is that it can be used to make comparisons over space and time possible, improving insights into core project characteristics like additionality and evaluating proposed carbon stock baselines. Given the diversity in scope, scale, and location of existing and emerging forest carbon projects, these remote sensing technologies could be instrumental in standardizing how we estimate carbon stocks. 

Yet, remote sensing adoption for MMRV is not without its challenges. Remote sensing approaches and geospatial analysis are complex, rapidly evolving, and relatively unstandardized for use in MMRV. To scale innovation and empower users, the MMRV community needs targeted tech transfer and clear guidance from scientific voices about how to use new tools and approaches responsibly and effectively.

Actionable guidance for using AI and remote sensing

To help bridge this gap for stakeholders, and particularly project developers, Meta commissioned Carbon Direct to evaluate the current state of remote sensing in forest MMRV. We developed two reports to provide strategic and tactical guidance:

• Remote Sensing for Forest Carbon: Barriers, Opportunities, and a Path Forward, undertakes a landscape view of remote sensing tools and datasets, explains barriers that exist in the current MMRV system, and makes recommendations as to how the remote sensing community might work together to address them. 

• Integrating Meta’s Canopy Height Map into Forest Carbon Methodologies: A Tactical Guidebook, explores how and in which circumstances project developers could leverage Meta’s open-source Canopy Height Map and model (called Every Tree Counts) to deepen their MMRV capabilities.  

Both reports are designed to objectively assess remote sensing technologies relative to forest carbon MMRV, providing project developers and other stakeholders with practical solutions to navigate the changing MMRV landscape and drive responsible adoption.

Meta’s open-source canopy height map: What it is and how to use it 

Canopy height maps (CHMs) are datasets used to estimate the height of forest vegetation, a key attribute for estimating carbon stocks. Meta has leveraged cutting-edge artificial intelligence to create an open-source model capable of combining high-resolution satellite and aerial imagery with lidar data. This model has been used to produce a high resolution (less than 1 m per pixel) CHM that estimates canopy heights for the entire planet based on data from 2018–2020. This open-source model can also be applied to additional satellite imagery to estimate canopy height for other time periods.