We Need the Largest Experiments on Earth to Achieve Our Climate Targets

Abstract

Natural Climate Solutions (NCS) are the climate mitigation sector with the largest potential, yet the least investment. A major barrier to scaling NCS is lack of evidence‐based scientific learning at the scale of actual interventions. To overcome this gap, we propose the world’s largest field experiments, supported by a global data‐sharing platform and guided by a clear framework of ethical principles and guidelines for delivering robust NCS evidence. This involves consistently reporting outcomes in units of MgCO2e $‐1 yr‐1 while describing direct and enabling actor groups and their financial, policy, and information interactions.

1. Introduction

Decarbonizing energy and industry is essential but alone will not achieve global climate targets. Natural Climate Solutions (NCS), such as climate‐smart forestry and agriculture, avoided deforestation, and reforestation, are described as the sector with the largest climate mitigation potential while contributing to a multitude of sustainable development goals (SDGs). Yet NCS (also reported as “AFOLU”) remains the least‐financed sector (IPCC ²⁰²²). Here we posit this is linked to a lack of advanced scientific learning from actual NCS interventions clarifying how and where their theoretical potential can be scaled ethically to deliver the biggest bang for the buck while managing risks. NCS projects are bespoke, and almost none are designed for evidence‐based learning.

Scientific learning is critical to delivering effective climate solutions, as it is for other global challenges like pandemic vaccines, famine, and poverty. In 1948, the publication of the first randomized controlled trial (RCT) ushered in antibiotics and the era of modern medicine (Crofton and Mitchison 1948). Scientific learning delivered COVID‐19 vaccines in record time, advanced the agricultural Green Revolution, and guided poverty reduction strategies around the world. This enabled the doubling of human life expectancy over the past century, tripling of crop yields since 1960, and the four‐fold decline in extreme poverty since 1980. Experimental methods have long been used in the development and manufacturing of technologies, including solar panels and electric vehicles. While experiments have proven that plants sequester carbon, we lack evidence on the biophysical and socioeconomic outcomes of alternative large‐scale interventions to reduce and reverse greenhouse gas emissions through NCS. That is, we need experimental approaches, and more broadly “prospective” impact evaluations, to transform NCS as they have for medicine and other fields like development economics (Duflo et al. 2007).

Each of the breakthroughs above is linked to evidence‐based scientific learning, coupled with a pressing need, explicitly measurable outcomes, rapid innovation, and both public and private sector investments. With NCS, the latter conditions are present, but the use of evidence‐based scientific learning at the scale of interventions is lacking—for both climate and other environmental program implementation. We are aware of only a few NCS interventions that used experimental or quasi‐experimental designs to determine CO₂e outcome per dollar of intervention (e.g., Jack and Jayachandran ²⁰¹⁸). Hence, we lack sufficient scientific evidence to identify, de‐risk, and cost‐constrain NCS interventions to deliver the speed and global scale required.

2. The Question

If NCS are to feasibly deliver a major contribution to climate stabilization, the central question that needs innovation and evidence to answer is: “How much climate benefit do NCS interventions deliver per dollar spent, and why?” The measurable and globally comparable outcome for answering this question is cost per tonne of additional CO ₂ e mitigation per year (MgCO₂e $⁻¹ year⁻¹). MgCO₂e refers to metric tonnes of carbon dioxide equivalents, a unit that standardizes the global warming potential of different greenhouse gases (and ideally also local biophysical effects like albedo).

This index integrates two values shared by all societies (climate stability and economic wellbeing) using measurable units that circulate globally. By normalizing the climate benefit relative to the economic cost, this index allows comparisons across scales, geographies, and interventions. This index annualizes CO₂e fluxes and associated costs and requires specifying both the discount rates and time periods involved.

This index must differentiate “additional” fluxes from existing ecosystem fluxes associated with business‐as‐usual land use. Establishing NCS “additionality” has been challenging; however, ongoing improvements to carbon verification standards, such as those incorporating dynamic baselines (e.g., VM0047 and VM0045), are addressing this challenge. Even as they improve, carbon verification standards are not designed or intended to report costs or answer questions about differential intervention effectiveness, so they support but do not answer our question above.

Innovation and evidence for delivering MgCO₂e $⁻¹ year⁻¹ is needed across spatial scales at which interventions occur, from trees to nations, and across fields including ecology, silviculture, agronomy, economics, political science, psychology, sociology, and engineering. Engineering may seem antithetical to natural solutions; however, many NCS are enabled by technology. These include (i) technologies that physically protect and restore ecosystem fluxes (e.g., precision agriculture to reduce fertilizer application, low impact forestry machinery), (ii) technologies that monitor, report, and verify outcomes, and (iii) technologies for distributing incentives to stakeholders (e.g., microfinancing cell phone software linked to monitoring systems).

3. Breadth and Depth of Evidence Needed

We need two transformations to answer the question above: (1) a breadth of consistent data collection and disclosure, especially on costs, and (2) a depth of evidence to determine why and where some interventions perform better than others, based on causal impact evaluations.

A breadth of evidence on MgCO₂e $⁻¹ year⁻¹ outcomes depends on both transparency in reporting and consistency in describing NCS interventions. Existing biophysical descriptions of ~20 types of NCS (Griscom et al. 2017) are necessary but insufficient for describing interventions. NCS interventions involve biophysical, financial, political, and technological interactions among many actors. While many parameters are important to describe an NCS intervention, at minimum four types of parameters should be consistently described (Figure 1): (1) “direct actors” who directly steward lands and waters, (2) additional biophysical changes, measured in CO₂e fluxes, due to changes in land and water stewardship, (3) “enabling actors” necessary for this change to be sustained, and (4) financial and other necessary interactions between and among actors for the additional CO₂e fluxes to occur, distinguishing transfers from true costs.

FIGURE 1.

Four types of parameters represented here are needed to consistently describe NCS interventions and their outcomes, both across a breadth of interventions and when interventions are designed as large‐scale experiments. Examples of categories that could be used to describe direct and enabling actor groups, and their interactions, are depicted. Costs incurred by direct and enabling actors and not reflected by transfers between these groups are also critical to record as part of the MgCO₂e $⁻¹ year⁻¹ unit discussed here.

This breadth of evidence is essential but insufficient. We also need a depth of evidence from NCS interventions implemented as parallel treatments, alongside controls, to understand why differential MgCO₂e $⁻¹ year⁻¹ outcomes emerge from different NCS interventions. In fields ranging from medicine to economics, RCTs are considered by many as the gold standard of “prospective” impact evaluation (Gertler et al. 2016). However, idealized randomized designs are not easy at large scales, and we thus need careful prioritization of the most critical questions that can be answered with ethically designed experiments. For example, one could design interventions with different contract lengths to assess trade‐offs between length of interventions and uptake by smallholders.

The question, the context, and the explicit theory of change should determine the study design, rather than constraining questions to fit an idealized randomized design. For some scales and some types of interventions, a variety of quasi‐experimental methods are better suited to help identify solutions for climate, people, and biodiversity that work as well as possible at scale (Ferraro and Pattanayak 2006).

4. A Call to Action

To achieve this breadth and depth of NCS evidence at a speed and scale to match the urgency of climate change, we propose three parallel initiatives: (1) Principles and Guidelines for ethically delivering the breadth and depth of NCS evidence discussed here, (2) an NCS Data Platform for sharing MgCO₂e $⁻¹ year⁻¹ and other outcomes crowdsourced by private and public project and program managers, and (3) NCS Acceleration Labs for launching NCS interventions as large‐scale experiments.

The first two will build on principles and guidelines and data sharing platforms already developed in other fields (e.g., World Medical Association ²⁰²⁴) and would begin by convening a broader group of scientists, practitioners, and other stakeholders including Indigenous Peoples and Local Community representatives. Aligned with the principles and guidelines and learning from evidence synthesis enabled by the data platform, the NCS Acceleration Labs would launch the largest set of experiments, and quasi‐experiments, on Earth to identify and refine low‐cost and low‐risk NCS interventions best positioned for gigatonne scaling. While the Large Hadron Collider may be the largest engineering experiment with a footprint of about 6000 ha, the largest experiment we know of based on area and mass is ecological, occurring across 110,000 ha of the Olympic Experimental State Forest in Washington, USA (Washington Department of Natural Resources ²⁰²⁵).

NCS projects and programs are already underway at similar and larger scales; however, they remain a series of bespoke interventions, not yet structured as landscape experiments to accelerate learning. Existing NCS projects and programs are significant yet remain tiny in the context of the > 300 billion dollar NCS investment gap (IPCC ²⁰²²) and the > 10 GtCO₂e annual mitigation potential of NCS (Griscom et al. 2017). Globally coordinated NCS interventions structured as ethical experiments, at scales considerably larger than 110,000 ha, are urgently needed to deliver evidence about the relative performance—in MgCO₂e $⁻¹ year⁻¹ and metrics of human well‐being and biodiversity—across a range of NCS intervention designs and geographies. By anticipating project failure, as both unavoidable and an opportunity to learn, NCS Labs would demonstrate the virtuous cycle of rapid scientific learning that has been demonstrated in other fields. As such, and at a comparable cost to the $4 billion Large Hadron Collider, NCS Acceleration Labs would implement the largest—in hectares and mass—experiments on Earth to stabilize our biosphere.

At the foundation of these advances is a shift in mindset. The chess grandmaster, Gary Kasparov, attributed his success to a “system mindset” where each game is played to better understand how to win, rather than just to win (Bahcall 2019). We are playing games with our biosphere and losing. Achieving a system mindset for NCS will require innovative experiments, financing, radical transparency, respect for indigenous and local communities, and widespread collaboration to rapidly deliver gigatonne‐scaled, ecologically responsible, socially just, and cost‐feasible Natural Climate Solutions.

Author Contributions

Griscom wrote the original draft following conversations with multiple co‐authors. All authors reviewed and edited it to reach the final manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Griscom, B. W. , Cook‐Patton S. C., Sills E., et al. 2025. “We Need the Largest Experiments on Earth to Achieve Our Climate Targets.” Global Change Biology 31, no. 10: e70555. 10.1111/gcb.70555.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.