Carbon markets are gaining steam. Carbon offset projects are often tied to projects that plant trees, protect forests, or otherwise use land-based systems to try to mitigate the climate crisis. These “natural climate solutions” (NCS) generally aim to preserve carbon stocks in living landscapes (e.g., protecting old-growth forests) or enhance carbon sequestration through vegetation growth (e.g., planting trees). These have become popular strategies through which nations and corporations propose to meet climate goals. For example, nearly 96% of nationally determined contributions to the Paris Agreement include NCS as a climate mitigation strategy. Relative to other approaches, NCS are shovel-ready, cost-effective, and can be recognized by carbon market accrediting bodies. In the private sector, companies with net-zero commitments increasingly include NCS in their strategies. 

In addition to these strategic commitments, implementation and interest in NCS is also increasing in scale and scope. Investments in NCS have doubled in the last six years and conversations on social media have tripled. At the same time, while increasing ambitions to mitigate climate change are admirable, NCS commitments by governments alone already exceed 1.2 billion hectares, an area nearly four times the size of India and equivalent to the total global cropland area. As a result, critical questions around the equitable and fair deployment of NCS projects have emerged: What does this enthusiasm mean for people living in and around NCS projects? Do different types of NCS have better or worse impacts on human well-being and biodiversity? 

The short answer to these questions about NCS is, we currently do not know. A significant barrier to finding any evidence on human well-being and biodiversity impacts of NCS is that much of the relevant research to understand these impacts precedes the concept of “natural climate solutions,” which was only coined in 2017. If we are to understand the collective evidence on impacts of NCS beyond carbon — what we call “co-impacts” — we must review hundreds of disciplines and methods of study. Finding relevant evidence using traditional methods for synthesizing evidence would take a team of hundreds of researchers thousands of hours. To overcome this, we created a process using cutting-edge machine learning (ML) techniques to analyze and categorize millions of papers.

Here are some of the key things we found:

• The pathways with the highest mitigation potential, such as avoided forest conversion and forest restoration, have a large amount of evidence for impacts/co-benefits to people and nature. This is a good thing!
• However, some of the pathways important for global mitigation – such as wetlands protection and restoration – have limited evidence, highlighting areas requiring additional effort to understand the fuller set of impacts.
• We additionally mapped countries’ mitigation potential from NCS against their levels of evidence. Countries with high mitigation potential and large amounts of evidence are “action” countries, where we are shovel-ready. Other countries that are important for mitigation but lack abundant evidence are places to prioritize for additional research efforts. Many countries that have great potential to deliver carbon benefits from NCS have comparatively little research on co-impacts, such as the Democratic Republic of Congo. These maps and detailed tables are in our paper and the SI.
• Implementing NCS could also offer major benefits for other important objectives such as protecting endangered species and achieving poverty alleviation goals.

We also learned some lessons.

• Methodologically, the pipeline we developed is replicable for other multidisciplinary topics, like climate resilience.
• Diverse teams that include engineering, computer science, ecology, social science, etc. are critical for successfully implementing such a data-driven approach. For example, our team included academic and non-academic partners, like Lexunit that provided critical engineering perspectives. While automation via machine learning can greatly streamline analyses, these processes still require expert knowledge to design robust, effective systems and review outputs.
• The promise of natural language processing continues to increase. Future analyses will be able to use even more sophisticated models to assess our dataset or others for important relationships between NCS actions, locations, cost, equity, and outcomes.
• Generally, the tool landscape is rapidly changing. New open-science platforms (e.g. OpenAlex), tools building on computational network analyses, and of course the growth of LLMs could help with creating more targeted and/or automated search results. However, there are still many unresolved concerns around best practices to implement these, but this is a growing area of fruitful research.
• Empirically, many NCS pathways and co-impacts are examined in tandem, indicating potentially exciting opportunities for synergies across NCS pathways and their connections to different sectors of society, but also the potential for difficult trade-offs between pathways (e.g., competing land use preferences).

Overall, NCS must be deployed equitably and effectively, and also provide local human well-being and biodiversity needs to be just, politically viable, and sustainable. To make responsible and high-integrity NCS investments, we must first know whether and how NCS projects can lead to co-benefits and trade-offs for people living in and around NCS projects. Our study takes the first step in advancing this understanding by uncovering probabilistically relevant evidence on NCS co-impacts, which is available to explore here. 

While this work tells us a great deal about the distribution and state of NCS co-impacts evidence, more is needed to understand the direction of impacts between NCS and human well-being (e.g., are avoided wetland impacts positively or negatively associated with economic development?) and biodiversity (e.g., are fire management practices increasing abundance?). This will require new models, tools, and methods to expedite evidence synthesis. 

This year’s COP16 and COP29 mark the latest rounds of multilateral talks to combat climate change and biodiversity loss. COP16 ended without governments agreeing on a plan for how to fund needed biodiversity conservation. COP29 started with the adoption of Article 6.4 to create a global carbon market, but also ultimately failed to meet any meaningful funding commitment. While these hinder low- and middle-income countries’ ability to act on low-carbon development trajectories and adaptation to adverse impacts of climate change that are already present, the frustration embedded in these failures also shows the enormous pent-up demand for NCS, which will only be unleashed when agreements get ironed out. Companies and countries alike are looking to NCS to provide mechanisms to transfer funds to local communities, as the overall growth in NCS is reflected in the growth and size of the global carbon markets whose revenue has exceeded $100 billion. 

Understanding local impacts is critical for restoring faith in the possibility of a robust global carbon market. We must leverage insights gained from efforts such as our global evidence map to inform dialogues around safeguards, reduce uncertainty, and ultimately deploy NCS projects that meet climate change mitigation goals while ensuring net benefits to local communities and ecosystems.