I’m a physicist by training. So when someone says, "you can’t measure X,” my knee-jerk reaction is “challenge accepted!”
In my current work, I often encounter misconceptions about soil carbon measurement. In this series, I’ll break down a few refrains that I hear, how I think about them, and how we’ve tackled these measurement questions at Indigo to deliver high-quality carbon impacts.
On its surface, this claim is demonstrably false. There are dozens, if not hundreds of peer-reviewed scientific studies measuring exactly this phenomenon. Also, many of the folks who state this misconception will also say that “we know that modern agriculture has caused the loss of huge amounts of soil carbon.” Well…isn’t measuring the loss due to changes in agricultural practice the same kind of measurement problem as measuring a gain?
However, when I dig deeper, I usually find a more nuanced concern: You cannot measure soil carbon changes from agriculture because of:
…variability in soil carbon across a field
…costliness of soil sampling in a large project
…field-scale accuracy of models
The idea is that measuring soil carbon on a single field with high certainty is incredibly expensive, and would require lots of direct samples, controls, etc. And you have to do that scaled linearly to get to a project of millions of acres. It certainly sounds challenging, but we measure other parameters with these kinds of properties on a daily basis.
Have you ever checked what temperature it is outside or in your house? Seems simple, right? You look at a thermometer and read a number. But do you know how it really works?
The temperature in a room results from the energy of air molecules bumping about throughout the space. The more energy in these particles (from radiation from a heat source like your furnace or the sun through a window), the more they move and bump into each other, imparting energy across the whole space the gas fills. That’s what we essentially perceive as temperature. Now to measure temperature, do I need to zoom in and measure the velocity of every single air molecule in the room? That would be pretty darn expensive. Thank goodness I can rely on a model, the kinetic theory of gases, and an instrument like a thermometer that uses the principle of thermal equilibrium coupled to something like an enclosed liquid that follows another model, like the thermal expansion of liquids to read out temperature using a calibrated and validated instrument.
Why do I keep calling things models? A model is an approximation of the emergent properties of nature based on observation and first principles. For temperature, we’re relying on a field called statistical mechanics and thermodynamics. Models help us measure at scale through inference what would be tremendously hard to do directly. But all models need to be treated with care. They are usually calibrated within a range. For example, a low, medium, and high “standard” is used to check that the instrument measures those three points accurately and then it’s presumed that it operates well between them. Note, you don’t typically calibrate your thermometer against every 1 ℃ increment.
Take a model based measurement tool out of its calibrated and validated use case, and things might not work. A standard, alcohol based liquid thermometer may work for most chemistry applications, but when I did a low temperature physics lab to measure the superfluid transition of helium (which occurs around -271 ℃), we needed a very different measurement tool.
Turning back to the soil carbon measurement challenge; yes, individual fields are quite variable but by aggregating many fields together and looking at the project level, it’s similar to looking at the gas rather than the molecules. With the right calibrated and validated tools, we can get a measure within a reasonable range of certainty. In this case, the model we use as our “thermometer” is based on biogeochemistry coupled to Bayesian statistics rather than thermodynamics and statistical mechanics. The premise is that the fundamental processes that link the biology of plants and soil life with the geology of the minerals in the soil and the chemical interactions that move matter between them can be approximated through a combination of observation and first principles. But just like we calibrate and validate our thermometer model, we calibrate and validate our models against the peer-reviewed and published studies. We have to make sure these studies cover a sufficient range of climate zones, of soil types, of the crop systems, and practices we operate in (like those low, medium, and high standards for the thermometer) so that we can feel confident it works across the range between those tested points. I’ll happily always take more validation points, but I’d rather have them in areas outside our current range than somewhere between two points we already have.
At Indigo, we’ve worked hard to make sure our models meet the validation guidance that was set by a group of experts as part of the Climate Action Reserve’s Soil Enrichment Protocol working group. While others have been granted some pretty big exceptions or written other approaches that open up a potential for bias and inaccuracy in projects, we’ve held ourselves to the standard as written to operate within the ranges where we have literature to validate our models.
So yes, we CAN measure the impacts of regenerative agriculture on soil organic carbon at scale. We do it using calibrated and validated models, just like we measure temperature with a thermometer. And we’ve published on our approach multiple times (1, 2).
(1) Brummitt, C.D., Mathers, C.A., Keating, R.A., O’Leary, K., Easter, M., Friedl, M.A., DuBuisson, M., Campbell, E.E., Pape, R., Peters, S.J.W., Kumar, A.A., 2024. Solutions and insights for agricultural monitoring, reporting, and verification (MRV) from three consecutive issuances of soil carbon credits. J. Environ. Manage. 369, 122284. https://doi.org/10.1016/j.jenvman.2024.122284.
(2) Mathers, C., Black, C.K., Segal, B.D., Gurung, R.B., Zhang, Y., Easter, M.J., Williams, S., Motew, M., Campbell, E.E., Brummitt, C.D., Paustian, K., Kumar, A.A., 2023. Validating DayCent-CR for cropland soil carbon offset reporting at a national scale. Geoderma 438, 116647. https://doi.org/10.1016/j.geoderma.2023.116647.