Through carbon credit and sustainable crop programs, Indigo Ag is providing financial incentives for agricultural systems and practices that are more sustainable (also known as climate smart agriculture). We’ve worked hard to create programs to align financial incentives with environmental impact so that outcomes are environmentally meaningful. We’ve done this by taking a science-first approach that leverages data, technology, and scale, while being transparent about uncertainty and managing it by working it directly into how we assess impacts.
Our Carbon by Indigo program that allows farmers to participate in voluntary carbon markets is one way that we are creating incentives to drive a planet positive outcome. Our approach involves:
- Grounding our programs on the collective scientific knowledge
- Quantifying uncertainty and building it into our incentives
- Sharing our methods openly and seeking feedback from the broader community
- Collaborating with others and advocating for and investing in research to advance the science behind sustainable agriculture
We’ve gone to great lengths to share our approach and work with third parties to validate and verify what we do. We’ve been able to scale our program and reward farmers. As activity and interest have increased around soil carbon as a climate solution, there have been questions and debates about scientific rigor and the role of soil carbon in carbon markets. Here we’ll identify a handful of important misconceptions and myths and address them head-on as a way to explain how we’ve set up our approach to conservatively account for soil carbon removals.
If you’re too busy to read through this whole blog post, here’s the TL;DR:d578.jpg?width=1650&height=2100&name=CARB-6Myths_chart%20(1).jpg)
Myth 1: Agriculture is not a significant part of the global climate solution.
The Intergovernmental Panel on Climate Change (IPCC) – an organization comprising multiple scientists and experts that also won the 2007 Nobel Prize – stated in their latest report:
“There is substantial mitigation and adaptation potential from options in agriculture, forestry and other land use, and in the oceans, that could be upscaled in the near term across most regions (high confidence)“ - IPCC AR6 Section 4.5.4
I’m a numbers person, so here’s a figure from the same report modified to focus on the agriculture, forestry, and other land use sector.
Note that the total potential of climate smart agriculture is the sum of the first two bars in that figure.
When approached with a systems perspective and methods to account for local conditions, climate smart agricultural practices can both reduce emissions and replenish soil carbon reserves. Not only that, but the data and tools exist today to link climate financing (such as carbon offset markets) with climate smart ag in a way that is scientifically rigorous and transparent.
To be clear, expanding the adoption of climate smart agriculture is not an excuse to avoid changes elsewhere. We need to decarbonize our economy. But we cannot have a complete, global solution without addressing agriculture because we must continue to feed the world and agriculture accounts for almost 20% of global emissions (FAO, 2020).
The question isn’t if we should support a sustainable transition of our agricultural system, but how. Similarly, carbon crediting is not sufficient to solve the climate crisis, but the fact remains that, when conducted with scientific integrity, it is a tool to drive immediate action that would not otherwise occur.
Myth 2: A single practice must have a big impact everywhere or we shouldn’t credit for it.
There are no silver bullets in sustainable farming. No-till and cover crops have different impacts in different places and cropping systems. Many meta-analyses have been done and while both of these practices may be generally beneficial, impacts on soil carbon vary significantly depending on the details (like soil type, management history, weather and climate, crop rotations, etc.). But variability alone is not all uncertainty. The right practice on the right location at the right time in the right system will make a positive impact. When these practices are done in the right places, they result in healthier soils, have agronomic benefits, and help farmers ensure the multi-generational stability of their operations. For example, PBS highlighted the benefits of no-till in certain areas in North Carolina, including some work on a long-term research site that one of our scientists, Cara Mathers, did for her PhD before joining our team.
A good measurement instrument should be able to reflect the impact of sustainable agriculture practices given the other variables. To understand this variability, we have assembled direct measurements with good controls from peer reviewed literature into a ground truth dataset as the basis for what and where we quantify soil carbon changes for our projects under Carbon by Indigo and Market+ Source. This dataset, as well as how we use it, gets reviewed by external experts and the registry we work with, the Climate Action Reserve. No cherry picking is allowed; the dataset includes locations where big impacts are seen, as well as sites where no impact or negative impacts on soil organic carbon are observed.
Below is a map of the sites with multi-year, published research data that we have used (Figure 1 from our Validation Report for DayCent-CR v1.0.2):
Myth 3: All uses of a model like DayCent are the same.
Biogeochemical models, including DayCent, are instruments that can be used in many different ways. Just like a musical instrument, the results you get depend on how you tune it, who is playing it, and the context in which they are playing. An untuned guitar in the hands of my toddler produces a very different result than a tuned guitar in the hands of a master like Carlos Santana. The quality of music from Yo-Yo Ma playing the cello solo is dependent on him while Yo-Yo Ma playing cello in an orchestra is dependent on all the other instruments around him.
DayCent is one instrument that Indigo uses amongst the orchestra of technologies that enable us to measure, monitor, verify, and report the GHG emissions and soil carbon sequestration of agricultural projects. Indigo uses a specific version of DayCent called DayCent-CR (more on DayCent-CR below) in concert with a variety of other instruments including data collection tools, remote sensing algorithms, 3rd party software integrations, and sophisticated data validation services to meet the requirements laid out in the Climate Action Reserve’s Soil Enrichment Protocol. DayCent has a long history, with its origins coming from scientists at Colorado State University. It’s also been used by researchers around the world as well as tuned and used by government agencies like the EPA and USDA for various inventories and tools. Each of these uses has their own context and specifications and is tuned to meet those needs and paired with other tools and instruments or interfaces to meet those goals. While these uses have demonstrated DayCent’s ability as a general instrument to cover numerous ecosystems and crops to advance our knowledge, none of them met the specific requirements for creating a high-quality carbon credit that are laid out in both the Soil Enrichment Protocol or Verra’s VCS VM0042 methodology. Specifically, these protocols require that models should be initialized with soil data that are representative of the project to ensure that it’s grounded in real world measurements that are actually relevant to the quantification of the project impacts.
DayCent-CR is a specific version of DayCent that has been developed to meet the rigorous standards of the latest soil carbon removal protocols. We’ve also calibrated, or tuned it, using a Bayesian approach - a specific statistical technique that enables fine tuning based on what we know from prior knowledge - and then validated it (demonstrated performance against that ground truth dataset we mentioned earlier) following the requirements of the Model Validation Guidance created by an expert working group with the Climate Action Reserve. This tuning ensures that the model more accurately conforms to the reality described by published experimental results in the validation datasets. Section 3.4 specifically addresses how Model Validation Reports are reviewed and approved by an external expert and Climate Action Reserve under a model very similar to the peer-review process of an academic paper. You can see our latest Model Validation report here. We’ll talk more about the Model Validation process in the next section.
Beyond the Model Validation, each of Indigo’s carbon harvests goes through a rigorous process which includes third-party auditing and submission of data. The company’s first submission included 17,000 documents of scientific and farmer data shared with the auditors, covering activities on over 100,000 acres.
The protocol that we follow for our US project (CAR1459), the Soil Enrichment Protocol, was developed with an expert working group that included key expert stakeholders in the sector, including research groups, professors, technical experts from the United States Department of Agriculture, multiple NGOs, verifiers, project developers (including Indigo), and farmers.
Myth 4: We need to have absolute certainty (i.e., zero uncertainty) to take action.
Farmers are continuously taking action in the face of uncertainty. They don’t control the weather, the price of commodities, or the cost of inputs, but they manage these uncertainties to produce the goods on which we depend. Farmers constantly account for variables both within and outside of their control and do their best to manage the noise of uncertainty.
On the science side, we face a similarly daunting number of variables and variance, so we have put considerable time and expertise into the development of scientific methods to manage this uncertainty. While we can explain a good amount of variability, agroecosystems are complex, and there are areas of variability that our current models don’t explain. But we can still quantify and conservatively account for the uncertainty, even if we cannot explain it.
The Model Validation Guidance sets out clear guardrails to ensure projects credit only in areas where there are robust published datasets AND where a model has demonstrated adequate performance. After calibration, we take that ground truth dataset and test our calibrated model against it. That means:
- We have to demonstrate that our 90% prediction intervals contain the observed measurements 90% of the time. This kind of benchmark is similar to those used by the Food and Drug Administration to evaluate the performance of home glucose tests for diabetics.
- We have to quantify the residual error and demonstrate that there is no significant bias by showing that the model bias is smaller in magnitude than the pooled standard error of the validation dataset, which is computed from replicate direct soil measurements.
- As written, the CAR SEP stipulates that we need to demonstrate the above for each combination of crop functional group and practice category in our project (i.e., just because we can model cover crops grown with corn in Iowa doesn’t mean we could automatically use the same approach for cover crops grown in the tropics).
The figure below shows what the modeled results look like compared to the experimental results for all the data together (Figure 18 from our Validation Report). The takeaway is that we have a strong fit between model predictions and the measured validation data. This is the value of only using the model where supported by robust validation datasets
All of this analysis has to be reviewed and confirmed by an outside expert and then reviewed and approved by the registry (Climate Action Reserve), with the details of that review made public alongside our calibration and validation report. Our calibration of DayCent-CR (a version of DayCent developed to and calibrated to meet the requirements of the Soil Enrichment Protocol) has passed all of these requirements. All approved model validation reports are published on the CAR SEP website, so interested experts are able to not only review the reports and the expert reviews, but also compare different reports to each other.
Notice, that the requirement is not that the model variance or uncertainty has to be zero. It has to be quantified and then it gets combined with other sources of uncertainty, such as sampling uncertainty, to come up with a total uncertainty for the project. But what if the uncertainty at a field scale is large?
I’m originally trained as a physicist and when I learned statistical mechanics, I learned the power of aggregation. Temperature is a function of the energy of the air molecules bouncing all around you. While it’s hard to know the energy of any individual air molecule with certainty, we’ve figured out how to measure a room’s temperature with pretty high confidence. Similar statistical concepts are used in other areas that we rely on every day like medicine and insurance. We take an aggregation approach by putting together large numbers of fields into a project to estimate the project level impact (similar to the room temperature). We then take this impact and apply a discount based on the total uncertainty for the project before the registry issues credits. This means that, statistically speaking, it's more likely that the actual climate impact is greater than the total credits issued.
What about uncertainties in data or uncertainty in how long practices will continue? First of all, we only issue credits on an ex-post basis, meaning we have monitored the activities on the fields and quantified the impacts. In addition, we make conservative choices along the way, we have plans for long term monitoring for permanence, and we contribute a portion of each issuance into a registry-managed buffer pool (based on a risk assessment). Details of all of these items (and much more) are explained in our Monitoring Plan for CAR1459, along with associated supporting documents. These are all available for download from the CAR registry, with new files posted following each successful credit issuance.
The result is a system that creates the highest value where the impact and scientific certainty is highest, and sets a threshold below which low impacts and high uncertainty don’t produce creditable results. This design prevents low quality credits and incentivizes continuous improvement of the entire system, including the sampling design, the data collection, and, yes, the biogeochemical models.
In our first carbon harvest, the total uncertainty deduction was 37%. We made improvements in our sampling operations, our statistical variance techniques, and improved our model calibration so that in our second carbon harvest, the total uncertainty deduction was 18%. Even given these improvements, based on the characteristics of new fields that enroll over time, uncertainty could go up, but we will always accurately account for it.
Myth 5: Choosing to do nothing or waiting to take action is better than acting in a state of uncertainty.
When considering an action, we have to consider the unintended consequences. But it’s also important to remember that choosing NOT to act is itself an action that has consequences. In this case, we already know the consequences of not incentivizing climate smart agriculture. We’ve lost between 115-154 billion tons of carbon from soils globally (Sanderman et al., 2017; Lal, 2018). Climate smart agricultural practices, when implemented with consideration for local history and conditions, have many co-benefits including reduced soil erosion, reduced runoff into waterways (Atwood and Wood, 2020; Keesstra et al., 2016), weed suppression, and reduced fuel use to name just a few. By creating economic incentives for climate smart agriculture, we create a vehicle to support rural economies as well. Section 2.3 of the CAR1459 Monitoring Plan goes into further detail on the sustainable development benefits of regenerative agriculture.
Myth 6: Private endeavors do not contribute to public advancement of science.
The United States has a proud history of science and technology flowing back and forth between the public and private sectors. Technologies we take for granted everyday, like Google search and Siri, came out of federally funded research, but it took significant investments from the private sector to turn them into technologies that could touch the lives of people around the world. Further, the use and evolution of the underlying technologies continue to support public sector programs and goals.
As a science-led company, we’re deeply engaged with the broader science community, learning from the best knowledge and expertise available, looking for feedback on our approaches, and sharing back what we’re learning and developing. Here are some of the ways that we are sharing our work to advance the broader space:
- By working with high-quality registries, we make many details about our projects, methods, and results public.
- For the past two years, we’ve hosted an in-person Science and Policy Forum to gather stakeholders from across academia, non-profits, other startups, and the farming community to have frank discussions about the state of the field and to discuss our approaches and open questions.
- We’ve previously shared how the advancements in the Soil Enrichment Protocol enable for a sophisticated and ever improving approach to agricultural soil carbon projects, including its treatment of model uncertainty, its hybrid approach to sampling and modeling, and its approach to dynamic and adaptive baselines to define “what would have happened” in a peer-reviewed article.
- We’ve been showing up at key conferences such as the Tri-Societies Conference (ASA, CSSA, and SSSA) and the American Geophysical Union Conference (AGU) to share findings from our research, learn from others, and gather feedback for improvement:
https://scisoc.confex.com/scisoc/2021am/prelim.cgi/Paper/133642
https://scisoc.confex.com/scisoc/2021am/prelim.cgi/Paper/134231
https://scisoc.confex.com/scisoc/2022am/meetingapp.cgi/Paper/145877
https://agu.confex.com/agu/fm22/meetingapp.cgi/Paper/1156041
https://agu.confex.com/agu/fm22/meetingapp.cgi/Paper/1181406
https://agu.confex.com/agu/fm22/meetingapp.cgi/Paper/1164825 - For areas where we find there is insufficient research, we invest, we partner, and we learn together. Along with multiple universities and other organizations in the Precision Sustainable Agriculture network, we are exploring ways to improve cover crop implementation. With multiple university partners, we are examining the relationship between grazing practices and soil organic carbon. Through a USDA-funded grant, we are partnering with the National Indian Carbon Coalition to help support adoption of climate smart agriculture among tribal producers.
Here’s one of our Central Indiana on-farm trials in action with cover crops on the left and no cover crops on the right:
While we take action now in areas where we have sufficient certainty and understand the variability, I’m excited about the areas we’ll be able to impact tomorrow through the work we are doing together with our collaborators to make scientifically meaningful incentives for climate smart agriculture available to all.
In Conclusion: Strong foundations in science are important for all ecosystem accounting
When evaluating the scientific rigor behind carbon credits, it’s also important to keep in mind how they are used. Corporations follow guidance, such as the Greenhouse Gas Protocol, to quantify their annual GHG footprint. Even a modest exploration of the data and methods used to develop these inventories can highlight significant data gaps and areas of uncertainty. This is especially true of “Scope 3” emissions, arising from sources in a company’s value chain outside of its direct control. All of these can benefit from science and methods that are constantly improving. We fully believe that carbon credits should be held to a higher standard than the minimums required for an inventory, and are proud of the fact that we have met this higher standard. But the result is that we have more confidence in the accuracy of the credits we create than in many corporate inventories. It is precisely because of the incentives created by carbon markets that we are able to devote the resources necessary to generate credits with high scientific rigor at scale. It also means we have the tools and science to support higher quality inventory accounting, which is also critical if we want to get incentives aligned and have meaningful outcomes as a planet. Through our Market+ Source program Indigo is bringing the same rigorous quantification methods we pioneered in carbon credits over into the world of agricultural supply chains.
We absolutely need to engage in meaningful debate as scientists and a society. Myths and misunderstandings can hold a field back from taking action; and we’re running out of time to take action. We need to overcome these misunderstandings so that we can surface and address valid concerns to move the field forward, protect the soil and our ability to feed the world, and address a rapidly changing climate. To accomplish this, we are committed to scientific transparency. We welcome questions and debate on how we can incentivize sustainable agriculture better. Our door is open. We invite you to partner with us; provide your critical assessment and constructive feedback, and help contribute to a planet positive future. Let’s advance where the ground is firm and work together to expand the areas where we can take actions.
References cited, listed alphabetically:
Atwood, L. W., & Wood, S. A. (2020). AgEvidence: Agro-Environmental Responses of Conservation Agricultural Practices in the US Midwest Published from 1980 to 2020. Knowledge Network for Biocomplexity.
Climate Action Reserve Requirements Guidance for Model Calibration Validation, Uncertainty, Verification for Soil Enrichment Products Version 1.1a. (2022). Available online at: https://www.climateactionreserve.org/wp-content/uploads/2020/10/SEP-Model-Requirements-and-Guidance-v1.0a.pdf
Climate Action Reserve Soil Enrichment Protocol Version 1.1. (2022). Available online at: https://www.climateactionreserve.org/wp-content/uploads/2022/06/Soil-Enrichment-Protocol-V_1.1-final.pdf
FAO. (2020). Emissions due to agriculture. Global, regional and country trends 2000–2018. FAOSTAT Analytical Brief Series, No 18. Available online at: https://www.fao.org/3/cb3808en/cb3808en.pdf
Indigo Ag. (2022). Validation Report DayCent-CR Version 1.0.2.
https://www.climateactionreserve.org/wp content/uploads/2022/11/CAR1459_model_val_DayCentCR_1.0.2.pdf
IPCC. (2023). Climate Change 2023: Synthesis Report. A Report of the Intergovernmental Panel on Climate Change. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, H. Lee and J. Romero (eds.)]. IPCC, Geneva, Switzerland, (in press)
Jackson Hammond, A. A., Motew, M., Brummitt, C. D., DuBuisson, M. L., Pinjuv, G., Harburg, D. V., Campbell, E. E., & Kumar, A. A. (2021). Implementing the Soil Enrichment Protocol at Scale: Opportunities for an Agricultural Carbon Market. Frontiers in Climate, 3. https://doi.org/10.3389/fclim.2021.686440
Keesstra, S. D., Bouma, J., Wallinga, J., Tittonell, P., Smith, P., Cerdà, A., ... & Fresco, L. O. (2016). Forum paper: The significance of soils and soil science towards realization of the UN sustainable development goals (SDGS). Soil Discussions, 2016, 1-28.
Lal, R. (2018). Digging deeper: A holistic perspective of factors affecting soil organic carbon sequestration in agroecosystems. Global Change Biology, 24(8), 3285–3301. https://doi.org/10.1111/gcb.14054
Sanderman, J., Hengl, T., & Fiske, G. J. (2017). Soil carbon debt of 12,000 years of human land use. Proceedings of the National Academy of Sciences of the United States of America, 114(36), 9575–9580. https://doi.org/10.1073/pnas.1706103114
Verra VM0042 Methodology for Improved Agricultural Land Management Version 2.0. (2023). Available online at: https://verra.org/wp-content/uploads/2023/05/VM0042-Improved-ALM-v2.0.pdf