The formula, in plain terms

VM0042 includes the standard power-analysis formula for this problem. The number of samples you need is:

Three quantities drive it:

• MDD is the minimum detectable difference: the size of SOC stock change you want to be able to detect (in Mg C/ha).
• S is the standard deviation of the change in SOC stock between your two sampling times.
• t_α and t_β set your statistical standards: the false-positive rate (α, usually 0.05) and the power to detect a real change (usually 90%).

Smaller changes are harder to detect, so a smaller MDD demands more samples. More variable soils are harder to measure, so a larger S demands more samples.

S is the variability of the change, not the variability of the soil

This is the single most important detail. S is not the spatial variability of soil carbon at one point in time. It is the standard deviation of the difference at each location between the first and second sampling.

That distinction is the entire reason to resample the same locations over time rather than drawing fresh random samples each period. If you sample independent locations each time, the difference between the two surveys carries the full spatial variability of the landscape at both times: a sandy knoll measured once and a depression measured the next, differences that have nothing to do with your practice. When you return to the same georeferenced points, those large, static spatial differences subtract out, because each location is its own control. What remains in S is the genuine variation in change plus measurement and relocation error.

However, it is important to remember that spatial variation is still a concern, as fine scale variation in SOC can drive significant differences at the scale of a few meters or less, which makes consistent resampling challenging.

In the Saskatchewan data, the spatial standard deviation of SOC stock is roughly 16 Mg C/ha, which would put the standard deviation of an independent difference near 22 Mg C/ha. Resampling the same sites brings the standard deviation of the change down to about 9.7 Mg C/ha. Pairing more than halves the standard deviation and cuts the variance by about 80%, and because S is squared in the formula, that is a large reduction in samples. This is why VM0042 requires resample locations to be georeferenced and sampled in the same season: to preserve that pairing.

What real variability looks like

Here is where grounding the numbers in real data matters, because it changes the conclusion. Using the paired standard deviation of the change from each dataset:

• Saskatchewan no-till (benchmark network, 0 to 30 cm): S ≈ 9.7 Mg C/ha. This is a province-wide network of no-till farms resampled over 22 years.
• Ontario cover cropping data (0 to 15 cm): S ≈ 4.3 Mg C/ha. This comes from a controlled field study with uniform management and tight relocation, with all plots under cover-crop treatments.

The gap between those two is consistent with what you would expect moving from a controlled experiment to a network of real farms, although the datasets also differ in region, soils, practice, and sampling depth, so it is not a clean like-for-like comparison. The practical lesson still holds: the variability you measure at a research site is likely to understate what you face across working fields, and operational projects should expect values closer to the higher number.

Stratifying does less than you might hope. Within individual Saskatchewan soil zones, the paired S still ranges from about 6 to 11 Mg C/ha, so splitting the network by zone does not change the picture much.

One more check matters before applying these numbers to shorter intervals. If S grew with the length of the interval, a 22-year S would overstate the noise for a 5-year measurement. In the Saskatchewan data it does not: the paired S is about 10.6 over 9 years, 8.6 over 13 years, and 9.7 over 22 years. The variability is dominated by measurement, relocation, and small-scale soil differences rather than by how long the change has been accumulating.

Now put those values against expected rates of change. The forward-looking rates, based on modelling, are small:

• No-till in Saskatchewan: about 2.3 Mg C/ha of additional soil carbon over the next 75 years, or roughly 0.03 Mg C/ha per year.
• Cover cropping in Ontario: an effect of about 0.045 Mg C/ha per year.

Over a five-year monitoring interval, those become changes of roughly 0.15 Mg C/ha (no-till) and 0.22 Mg C/ha (cover crop). Feed them into the formula with the real S values, and the sample numbers become quite large.

Detecting the five-year cover-crop effect takes on the order of 3,900 samples. Detecting the five-year no-till gain takes about 42,000. Sampling that much would be completely impractical.

These figures also assume the simplest design: measuring change at one set of sites. If a project instead compares project sites against baseline control sites, it has to detect a difference between two groups’ changes, which roughly doubles the samples needed in each group. By the standard of a statistical test, these changes cannot be confirmed by direct measurement over a five-year window.

Precision has a price in VM0042

VM0042 does not actually require a project to pass that test. The methodology states that projects “are not required to take” the number of samples the power analysis suggests. What enforces precision instead is the uncertainty deduction (VM0042 Equation 74). The deduction applied to a project’s credited reductions is about 0.43 times the relative standard error of the estimate: the noisier your estimate relative to the change itself, the larger the share of credits deducted. A noisy estimate is not rejected, but it does significantly reduce the total credit value.

Additionally, the risk of failing to detect a real change rises sharply at lower sample numbers, particularly when the true change is small and variability is large.

That raises the next question: not “can I detect the change?” but “how much of the change will I lose to uncertainty?” Applying the Equation 74 rule (the uncertainty deduction equation) to the same five-year changes and real S values:

These estimates are optimistic, because they count only the sampling uncertainty in the project change. Uncertainty in the baseline and in any model used adds to the deduction. A five-year no-till measurement earns essentially nothing at any realistic sample size: below about 740 samples, the deduction consumes the entire change. A cover-crop project with a few hundred samples would earn credits, but lose around half of them to uncertainty.

With few samples, there is also a real chance that the observed change comes out near zero, or even negative, purely by chance, leaving nothing to credit.

Waiting is the main practical lever

If sampling harder cannot rescue a five-year measurement, what can? Time. Because the cumulative change since the baseline grows as you wait while the variability stays roughly constant, the required samples fall roughly as one over the elapsed time squared. Every doubling of the time since baseline cuts the samples needed roughly fourfold.

However, VM0042 requires SOC stocks to be remeasured at least every five years. Proponents need to be realistic about what is likely to be observable over a five-year period: small changes, carrying large uncertainty deductions.

The figure shows both the power and the limit of this lever. The requirement drops steeply as the interval grows. But for these slow rates and real variances, reaching even a large program of around 100 samples takes roughly three decades for the Ontario cover-crop case, and more than 50 years for the Saskatchewan no-till case. Five- and ten-year measurement intervals do not get you there. The arithmetic only works over multi-decade baselines.

Sampling design can shrink the signal before you start

Power is not the only thing a measurement program controls. Design choices can change the effective size of the signal itself, and depth is the clearest example. SOC change from surface-acting practices concentrates near the surface, while deeper layers add carbon and variance but little signal. Compositing over a deeper interval therefore dilutes the change you are trying to detect.

The Saskatchewan data show this directly. The gain over 1996 to 2018 concentrates in the top 10 cm: the 0 to 10 cm layer holds about two-thirds of the 0 to 30 cm change, while carrying roughly half the variability.

Measuring the 0 to 10 cm increment needs about 149 samples to detect its change; integrating down to 0 to 30 cm raises that to about 251. The deeper you composite, the more the added variance outruns the added signal. VM0042 requires reporting to at least 30 cm on an equivalent-soil-mass basis, and it permits SOC content to be measured on a single mixed sample spanning 0 to 50 cm. That is convenient, but if the change lives in the top layer, a single deep composite buries it. Measuring and analyzing by increment, rather than bulking, preserves the signal that a deep composite averages away.

The takeaway

For the slow SOC gains that no-till and cover cropping actually deliver, direct measurement over a short interval cannot reliably detect the change, and under VM0042 the resulting uncertainty deduction will consume most or all of the credits. That leaves three options: wait decades for the signal to accumulate, use a well-validated model with periodic measurement to true it up, or accept a large uncertainty deduction on whatever you measure. VM0042 provides for all three, and the model-based pathway exists precisely because short-interval measurement cannot consistently detect SOC stock changes without an impractically large number of samples.

None of this is an argument against building soil carbon. The gains are real and worth pursuing. It is an argument for being realistic about what measurement can and cannot confirm on the timescales that projects operate on, and for designing monitoring (the interval, the depth, the pairing, and the stratification) around that reality.

Two pools, not one

Soil organic carbon (SOC) is the carbon held in organic matter: plant residues, roots, microbial products, and the stable organic carbon built up in soil over time. This is the pool that responds to management that adds organic carbon to the soil, and it is typically the pool that carbon programs are trying to build.

Soil inorganic carbon (SIC) is carbon held in carbonate minerals, essentially ground-up limestone in the soil. Its presence is driven largely by parent material, precipitation, and profile age. In dry and calcareous soils, including large parts of the Canadian Prairies and other semi-arid regions, the inorganic pool can be several times larger than the organic one. While SIC can also be influenced by management practices, the mechanisms that affect SIC and SOC are different. Detecting changes in either reliably requires isolating them with the right measurements.

Total carbon is simply the two added together:

Total carbon = SOC + SIC

The two pools form, move, and change for different reasons and on different timescales, and they are not distributed the same way through the soil profile. Organic carbon is concentrated near the surface and falls off with depth. Carbonate is the opposite: typically low at the surface and accumulates as you move deeper in the profile.

Treating these two pools as interchangeable, or collapsing them into a single “soil carbon” number, is where problems begin.

Why the measurement method matters

A standard laboratory dry combustion analysis measures all of the carbon in a sample at once: organic plus inorganic, reported as total carbon. In acidic soils that contain little or no carbonate, total carbon is essentially all organic, so this is fine. The complication arises in calcareous soils, where a substantial share of the carbon is carbonate, and measuring everything together no longer tells you the organic carbon.

And calcareous soils are not rare in the Canadian Prairies. Across agricultural Saskatchewan, a large share of soils are mapped as calcareous, concentrated through the central and eastern plains. These soils have significant SIC contents in the B horizon (near-surface subsoil), and sometimes the A horizon (the topsoil).

A calcareous classification is a direct indicator, but it does not catch everything. The Canadian soil classification has no calcareous Vertisol subgroup, so heavy clay Vertisols are never flagged as calcareous even when they carry carbonate. Soil pH captures these cases: alkaline soils are where carbonate, and therefore inorganic carbon, is most likely to be present. The calcareous classification and soil pH together give a good idea where inorganic carbon is likely to matter for soil carbon measurements.

Measuring SOC and accounting for SIC

Every method starts from the same reality: the instruments measure carbon, and on calcareous soils that means organic and inorganic carbon together. What the methods differ on is how they separate the two.

Dry combustion: the base measurement. Dry combustion elemental analysis is the standard way to measure soil carbon. The sample is heated in oxygen to a high temperature (roughly 950 to 1,050 °C), all of its carbon is converted to CO2, and the CO2 is measured. The result is total carbon. On acidic, carbonate-free soils that total is effectively all organic, so dry combustion gives SOC directly. On calcareous soils it does not, and one of the following is needed to recover the organic fraction.

Remove the carbonate first: acid fumigation or HCl washing. If the carbonates are destroyed before combustion, what remains to burn is organic carbon, so dry combustion then measures SOC directly. Two common approaches:

• Acid fumigation. Samples are exposed to concentrated hydrochloric acid vapour in a sealed chamber, which reacts with and removes carbonates in place before the sample is combusted. It is suitable for small samples used in elemental analysers and reduces the risk of acid removal of some of the SOC. However, incomplete removal of SIC can be a concern.
• Acid washing (direct acidification). Dilute HCl is added directly to the sample to remove the carbonates, after which the sample is dried and combusted. It is effective but more hands-on, and care is needed: washing can leach some soluble organic carbon, and dolomite reacts more slowly than calcite, so incomplete reaction is also a risk. This method can leave HCl residues that are corrosive to laboratory instruments.

Both of these methods are difficult to get working consistently, and require well-controlled procedures and skilled laboratory technicians.

Measure the carbonate separately: pressure calcimeter. Rather than removing the carbonate, you can measure it. In a pressure calcimeter the sample is reacted with HCl in a sealed vessel, and the CO2 released from the carbonate raises the pressure, which is measured and converted to carbonate content (CaCO3 equivalent), i.e., the inorganic carbon. Soil organic carbon is then obtained by difference: total carbon from dry combustion minus inorganic carbon from the calcimeter. It is simple, inexpensive, and robust, and it has the advantage of giving you the SIC value explicitly, which is the number you want to be able to track on its own.

Separate them in one run: temperature-ramping methods. Organic and inorganic carbon release their CO2 at different temperatures: organic carbon combusts at lower temperatures, while carbonates decompose only at higher ones. Temperature-ramped or stepped-combustion instruments exploit this by measuring the CO2 evolved in separate temperature windows, so a single analysis can yield both SOC and SIC with no chemical pretreatment. The convenience is real, but the temperature windows can overlap (some stable organic matter burns hot, and some carbonate decomposes cooler than expected), so these methods need careful calibration and can misassign carbon between the pools if applied without it.

A cheaper alternative to elemental analysis, loss-on-ignition (LOI), estimates organic matter from the weight a sample loses on heating. It is an estimate rather than a direct carbon measurement, and it is biased in high-clay soils because structural water is driven off along with organic matter. It has its uses, but it is not a substitute for a direct, carbonate-aware carbon measurement when the goal is to quantify SOC accurately. While LOI data may be suitable for agronomic decision making, the precision is much too low for soil carbon change quantification.

Whichever route you take, the goal is the same: report soil organic carbon and soil inorganic carbon as separate numbers, not a single total.

Why this matters for carbon accounting

The practices that most (but not all) carbon programs credit, such as reduced tillage, cover cropping, and improved rotations, act on the organic pool. Soil organic carbon is what those practices build, and it is what a project is claiming to have changed.

Soil inorganic carbon can change too, but usually for entirely separate reasons: irrigation, liming, acidification, and the dissolution or precipitation of carbonates, all of which operate on their own timescales and have little to do with the above management practices primarily influencing SOC. Inorganic carbon was long treated as inert, but recent work shows it is not necessarily stable either, particularly in drylands under a changing climate.

Put those two facts together and the risk becomes clear. If you track total carbon as your indicator of change, a shift in the carbonate pool (or uncertainty in measurement) can mask or exaggerate the change in organic carbon that you are actually trying to measure, in either direction. On a calcareous soil where SIC dwarfs SOC, even a modest change or laboratory measurement error in carbonate content can swamp the SOC signal entirely. The result is a baseline, and a credited change, that may reflect inorganic carbon dynamics, or measurement uncertainty, rather than the organic sequestration the project set out to achieve.

The takeaway

When evaluating change in soil carbon, measure and report soil organic carbon and soil inorganic carbon separately, rather than relying on total carbon. On acidic soils the distinction may not matter much. On calcareous soils, which cover a large share of the Canadian Prairies, it is essential, and getting it wrong propagates straight into the baseline and the credited change. It is a small methodological choice at sampling and lab time, and a large one by the time the numbers are used to estimate carbon stock changes.

The reported accuracy overstates real-world performance

The data descriptor reports strong accuracy: an out-of-bag coefficient of determination (R-squared) of 0.82 for soil organic carbon at all three topsoil depths, and 0.70 to 0.76 for bulk density, sand, and clay.

The problem is what “out-of-bag” means. Out-of-bag error is an internal estimate produced by a random forest model from its own training data, not an independent test. For soil data it tends to overstate real-world accuracy, for two reasons that both apply here. The training points are spatially clustered, concentrated in the agricultural south and collected largely between 1980 and 2000. Soil properties are spatially autocorrelated, as a result out-of-bag points are frequently very close to training points the model has effectively already seen. Nearby points leak information, and the resulting out-of-bag validation scores reflect performance in densely sampled areas rather than at genuinely independent locations across the country.

The gap between that internal estimate and independent validation is substantial. When the maps were compared against independent provincial soil datasets:

• In the agricultural regions of southern Ontario, validation against roughly 1,500 recent soil samples found essentially no correlation for organic carbon, pH, sand, silt, clay, or cation exchange capacity.
• In southwestern Quebec, a comparison against tens of thousands of samples found little agreement for soil organic carbon, with many mineral soils (which hold well under 10% carbon) predicted to contain far more.
• In Saskatchewan, independent validation reached an R-squared of approximately 0.25 for topsoil carbon, roughly half the accuracy of other published provincial-scale carbon mapping in the same region. The dataset used for validation in this case is from the same time period as the data used to create the Soil Landscape Grids of Canada.

An out-of-bag R-squared of 0.82 set against independent validation ranging from near zero to about 0.25 is the single clearest reason to treat the current accuracy claims with caution.

The artifacts point to a systemic cause: covariate choice

The problems are not limited to accuracy statistics. The maps contain systematic artifacts created due to the environmental covariates used to build them.

Predictive soil mapping works by learning statistical relationships between soil observations and environmental covariates, and then applying those relationships everywhere. The output can only be as good as the covariates it is built on. If a covariate contains an artificial pattern, the model reproduces it in the soil predictions. Two covariate problems are clearly visible in the 0 to 5 cm soil organic carbon results.

Terrain covariates derived from a contour-based elevation model. Many of the covariates (slope, curvature, topographic position, valley depth, topographic wetness index, landform classes, and others) are derived from a national digital elevation model. When a soil map built on those terrain derivatives shows nested, concentric lines like the two examples below, from different regions, it means the elevation model itself carries contour-line interpolation artifacts, a known characteristic of national elevation data interpolated from digitized topographic contours. The terrain derivatives inherit the pattern, and the model recreates it straight into the carbon predictions.

Vector datasets used as covariates and masks. The covariate set also includes polygon (vector) data, such as a national surficial geology map, and the product applies a depth-to-bedrock layer compiled from multiple raster and vector sources. The authors themselves note that this compilation produced abrupt division lines that needed correcting. Hard, straight edges of exactly this kind appear in the output, including the discontinuity along the Manitoba/Ontario border below. Soil does not change character along an administrative or map-sheet boundary, so these edges are covariate or data-compilation artifacts, not real soil variation.

These are not pixel-level inaccuracies to be smoothed over. They are structural, methodological artifacts that trace back to covariate selection and preparation, and they are exactly the kind of issue that consultation with the predictive soil mapping community would catch before release.

The correct resolution depends on the training data

The product is mapped at 100 m. But the training data cannot support that resolution everywhere. The pedon data are clustered and were collected decades ago, so across most of the country the density of real observations behind any given pixel is very low. The historical pedon records also carry location uncertainties of roughly 300 m or more, which means the training points cannot reliably be tied to a 100 m pixel (the mapping resolution) in the first place.

Mapping at a fine resolution over a huge area with sparse, imprecisely located data tends to compress the real variation in soil properties toward the average, so the map can look detailed while representing the landscape poorly. The resolution of a predictive soil map has to be selected carefully based on the properties of the training data: the right resolution follows from the location accuracy and density of the observations, and should be decided on that basis rather than set to match the finest available covariate grid. A product mapped finer than its data can support will look more precise than it actually is.

How this happened

National-scale soil mapping is genuinely hard, and it is nearly impossible to do well without the data and expertise that are spread across the country. In Canada, soil science capacity is decentralized: much of the current soil data and predictive mapping know-how now resides in provincial governments and academia. That community was not engaged to provide feedback or contribute to the development of the Soil Landscape Grids of Canada, and many practitioners only became aware of the product after it was released. That gap in collaboration is the most likely root cause of the issues above, including covariate choices and validation decisions that external review would have flagged immediately.

The key point for anyone considering using it

The independent evaluation of the Soil Landscape Grids of Canada suggests that this product should be viewed as still in development and not ready for carbon accounting, modelling, or use for decision making. Independent validation is weak, the reported accuracy relies on out-of-bag error that overstates performance, systematic artifacts are present, and the product itself is still labelled as under evaluation and review. The authors also note that accuracy is expected to improve as more ground data are added, which is a fair description of a product still in development.

In its current form, the Soil Landscape Grids of Canada is best viewed as a preliminary soil map, released for review and revision with the Canadian soil mapping community. It is not ready to be used as an input for carbon accounting, biogeochemical modelling, or land-management and policy decisions. Using it for those purposes today would risk carrying its errors, and its artifacts, directly into downstream results.

A constructive path forward

None of this diminishes the value of the goal. A reliable national soil data infrastructure is worth building, and CanSIS has an important mandate in that space. A few steps would help this product get there:

1. Continue to support the Canadian Soil Data Portal initiative, so that the current, consolidated data required for credible national-scale mapping is actually available to draw on.
2. Continue developing the product in collaboration with soil scientists from provincial governments and the academic community, rather than in isolation. For example, the Ontario Ministry of Agriculture, Food and Agribusiness has world-leading predictive soil mapping initiatives and operational processes that could be built on. Extensive research in predictive soil mapping is also taking place at universities across Canada.
3. Validate independently and consult before release. Future versions should be tested against independent, spatially representative data (not out-of-bag error alone), with the results and the covariate choices reviewed with Canadian stakeholders before public release.

The expertise needed to make this product world class already exists in Canada. Bringing soil scientists across the country together is the path forward to build a national soil map that people can genuinely rely on, and many in the community would welcome the chance to help build it.

Sources: Geng, X., He, J., Grima, V., Jiang, Y., Tetreau, M., Crittenden, S., Kiley, S., VandenBygaart, A. J., & Vanrobaeys, J. (2025). 100 m soil landscape grids of Canada. Scientific Data, 12, 1178. https://doi.org/10.1038/s41597-025-05460-4. The independent provincial validation results summarized here were compiled by members of the Canadian soil mapping community. All figures were rendered by the author from the released product.

A decade of data from one farm

The analysis drew on 3,861 agronomic soil samples collected from the top 15 cm across roughly 4,700 hectares of a southeastern Saskatchewan farm, over 2013 to 2022. I paired the soil organic matter record with total annual precipitation from the nearest weather station, and compared five-year moving averages of each to look at the direction both were trending rather than the noise of any single year.

The two move together

The result was striking in its simplicity.

Soil organic matter and precipitation were highly correlated (r = 0.91). The relatively wet years through 2017 coincided with higher soil organic matter; the dry stretch that followed brought a clear decline, bottoming out around 2021 when both precipitation and soil organic matter hit their lowest point. Temperature, by contrast, showed only a weak relationship in this dataset, rainfall was the dominant climate lever.

The likely mechanism is straightforward: in drier years, crops produce less biomass, so less carbon returns to the soil as residue and roots. That reduction in carbon inputs appears to matter more than any change in decomposition rate, and it shows up quickly in the more labile fraction of soil organic matter.

Why this is a monitoring problem, not just a curiosity

Here’s where it matters for carbon markets and MMRV. If soil carbon can swing by this much in response to a few wet or dry years, then a monitoring window that happens to fall in a dry stretch can show a “loss” that has nothing to do with management and one that lands in a wet stretch can overstate the gain. The same field, honestly measured, can look like a success or a failure depending purely on when you sampled relative to the rainfall cycle.

Concretely: a project could bank a real management gain over its first five years, then post an apparent decline over the next five simply because of below average rainfall even as the underlying, management-driven trend stays positive. Read either window in isolation and you’d draw the wrong conclusion.

The implication isn’t that soil carbon monitoring is unreliable, it’s that it has to be done with the climate signal explicitly in view. Credible measurement, reporting, and verification need to separate the weather-driven swing from the management-driven trend, whether by normalizing for precipitation, using climate-matched baselines, or monitoring over windows long enough for short-term cycles to average out. Attributing a rainfall-driven change to a farming practice, in either direction, is one of the easier ways to get a soil carbon number wrong.

The takeaway

Long-term soil carbon gains from good management are real and worth crediting. But over the few-year horizons that monitoring programs actually operate on, precipitation can dominate the signal. Any soil carbon assessment that doesn’t account for it is at risk of measuring the weather and calling it management.

This article summarizes: Sorenson, P. T., & Bedard-Haughn, A. K. (2025). Soil organic matter is responsive to short-term climate cycles. Canadian Journal of Soil Science. https://doi.org/10.1139/cjss-2024-0103

Pairing soil maps with a carbon process model

Estimating future carbon change means simulating the system, not just extrapolating a trend. So this work links two things: predictive soil maps that supply the starting soil carbon, texture, and bulk density for every quarter section, and the CENTURY biogeochemical process model, which simulates how carbon actually cycles through the soil year by year under a given climate and set of crop residue inputs.

A couple of details make the estimate credible. I esetimated when no-till was adopted on each parcel using satellite imagery based on the disappearance of bare soil in satellite imagery so the model starts adding no-till carbon at the right time and place. I then validated it against the Prairie Soil Carbon Balance network, a set of long-term monitored benchmark sites; using those measured starting values, the modelled and observed carbon stocks agreed closely (R-squared of 0.80). I then ran it forward under three climate pathways - low, moderate, and high emissions - to 2100.

What no-till has done, and what’s left

This figure is the whole story in one line. The steep black curve is the historical change in soil organic carbon; the coloured curves are the three future climate scenarios.

Historically, no-till added about 2.8 Mg of carbon per hectare in the top 20 cm, the steep part of the climb. The forward projection carries two messages at once. Carbon keeps accumulating in every scenario, so no-till stays a net benefit. But the curve visibly flattens: the annual rate of gain falls from about 0.06 Mg/ha/yr today to under 0.02 Mg/ha/yr by late century. The soils are approaching a new equilibrium, and warming accelerates the decomposition that works against accumulation.

Scenario matters, but not the way you might expect. Under low emissions, cropland gains about 3.1 Mg/ha by 2100; under both moderate and high emissions, only about 2.3 Mg/ha and the two upper pathways nearly converge. The big step down happens moving from low to moderate warming, not from moderate to high.

In whole-province terms, no-till sequestered on the order of 6 million tonnes of CO2-equivalent per year in Saskatchewan as of the mid-2010s. That annual rate is projected to fall by roughly half or more by 2100 - to somewhere between about 1.2 and 1.8 Mt CO2e per year, depending on how much the climate warms.

The part that matters for carbon programs

Two findings should shape how anyone approaches soil-carbon crediting here.

First, the marginal tonne is getting harder to add. Cumulative stocks are still rising, but the rate is in decline. For crediting, that’s the crux of the additionality and baseline question: a project counting on future no-till gains is working against a shrinking per-year margin, and the forecast needs to reflect that rather than assume past rates continue.

Second, the remaining potential is not spread evenly. Much of the Black soil zone in the north has already realized most of its no-till gains, so its future potential is lower. The largest remaining upside sits in the drier Dark Brown and Brown soil zones of south-central and western Saskatchewan, which started lower and have more room to add carbon. For siting projects or targeting incentives, geography is decisive and a provincial average would mislead you in both directions.

Extending the gains further will likely take more than no-till alone: better residue management, and crop varieties bred for larger, more stable belowground inputs.

Why this kind of modelling matters

Turning a soil-carbon question into a defensible number means combining spatial soil data, a process model that respects the biogeochemistry, honest treatment of climate scenarios, and validation against real monitored sites.

This article summarizes: Sorenson, P. T., Bedard-Haughn, A. K., & St. Luce, M. (2024). Combining predictive soil mapping and process models to estimate future carbon sequestration potential under no-till. Canadian Journal of Soil Science. https://doi.org/10.1139/cjss-2024-0046

Building a forward-looking yield model

The approach pairs two things I work with often: climate projections and predictive soil maps. I trained a machine-learning model (a random forest) on two decades of crop-insurance yield records, using both soil properties (organic carbon, clay, and sand from earlier mapping work) and the climate variables canola responds to: July maximum and minimum temperature, the number of days above 30°C, the number of nights above 16°C, and annual precipitation.

Then I ran that model forward using 26 downscaled global climate models, across three emissions pathways: low (SSP1-2.6), moderate (SSP2-4.5), and high (SSP5-8.5). For every quarter section in the province, from 2025 to 2100. Using 26 models rather than one is the point: the spread between them is the uncertainty, and averaging across them gives a more honest central estimate than any single model could. The model validated reasonably against held-out data (R-squared of 0.51, RMSE of 429 kg/ha), in line with other Canadian canola yield models.

Temperature is the lever

The clearest signal was that canola yields track temperature more than precipitation. Modelled yields rose with temperature up to a point and then decreased once mean daily July maximum temperatures passed about 28.5°C. That threshold matters because it is exactly the range future summers push into and it means the yield response to warming is not gentle and linear but flips from modest to damaging once the heat crosses that line.

How bad, and where

This is the figure that pulls it together: the relative change in canola yield across all three scenarios (rows) and three future periods (columns), all measured against the near-term low-emissions baseline.

Read it top-left to bottom-right and the risk gradient is unmistakable. Under the low-emissions pathway (top row), yields stay close to baseline through 2100. Under the high-emissions pathway (bottom row), the map turns progressively red, with the most severe losses concentrated in the late-century, high-emissions corner.

The province-wide median declines in Canola yield, long term, came out to:

• Low emissions (SSP1-2.6): about -39 kg/ha (2%) - essentially stable
• Moderate emissions (SSP2-4.5): about -137 kg/ha (6%)
• High emissions (SSP5-8.5): about -349 kg/ha (15%)

Two findings matter as much as the headline averages. First, the losses are not spread evenly. The largest declines fall in the Black soil zone. Today’s most productive canola producing region because those higher-yielding regions currently see little heat stress and have the most to lose as summers warm. Second, the losses arrive sooner than end-of-century. Under the moderate and high pathways, meaningful declines show up by the medium term (2050-2075), not just at 2100.

I’d flag one honest caveat: because of how the model and the insurance-claim yield data behave at the extremes, these are likely conservative estimates. Real losses in the hottest areas could run higher than shown.

Why this kind of analysis matters

For anyone assessing physical climate risk, lenders, insurers, agribusiness, or producers planning decades ahead, the useful output is not a single number but a spatial, scenario-dependent picture: how exposure changes by location, by emissions pathway, and over time. That is what lets you separate a manageable risk from a structural one, and target adaptations, such as heat-tolerant varieties and earlier seeding to move flowering ahead of peak July heat, where it will matter most.

It also shows what pairing climate ensembles with predictive soil mapping can do: turn broad warming projections into decision-grade, field-scale risk maps for a specific crop and a specific place.

This article summarizes: Sorenson, P. T., Mood, B., & Shirtliffe, S. (2025). Future canola yields under different climate scenarios in Saskatchewan, Canada. FACETS. https://doi.org/10.1139/facets-2025-0121

Mapping the carbon first

You cannot study something you cannot measure everywhere, so the first step was a wall-to-wall map of topsoil organic carbon. I combined 5014 historical soil samples with satellite imagery, climate data, and terrain information, and trained a machine-learning model (a random forest) to predict SOC across the whole agricultural region at 250 m resolution. The result is the map below.

The model held up against independent test data (R-squared of 0.48, concordance correlation of 0.67). Solid performance for regional soil mapping built largely from legacy samples and open satellite archives.

Linking carbon to yield

With the map in hand, I paired it against crop insurance yield records - between 100,000 and 200,000 records per crop - and used a mixed- effects model that normalized both carbon and yield within each rural municipality. That normalization is the important part: it strips out the broad climate gradients that drive both SOC and yield in the same direction, so what remains is much closer to the effect of carbon itself rather than the effect of a wetter or cooler region.

The relationship was clear and positive for all four crops. A one percentage point increase in SOC was associated with, on average:

• Wheat: +263 kg/ha (about 7.9%)
• Barley: +293 kg/ha (about 8.2%)
• Canola: +133 kg/ha (about 5.6%)
• Field peas: +135 kg/ha (about 5.5%)

At the commodity prices used in the study, that works out to roughly $4,000 to $5,200 in additional revenue per quarter section for every one-point gain in SOC, depending on the crop. The cereals responded most strongly; peas the least, which fits their biology, as a legume fixing its own nitrogen, peas are less dependent on the nitrogen that richer soils supply.

Where it pays the most

The headline averages hide the most useful finding for anyone deciding where to invest in soil health: the returns are largest on the poorest soils. Yield response to increasing carbon contents was strongest in areas with lower starting SOC (roughly the 2 to 2.5% range) and tapered off as soils got richer, with the response flattening above SOC contents of about 3%. In other words, the same tonne of carbon buys more yield on a depleted soil than on an already-carbon-rich one.

Why this matters beyond the farm gate

The main takeaway from this work is that while much of the conversation on practices that build soil carbon (reduced tillage, cover cropping, better rotations) have been focused on sequestration, there is a real yield benefit from increasing SOC.

This article summarizes: Sorenson, P. T., Shirtliffe, S., & Bedard-Haughn, A. K. (2024). Examining the effect of soil organic carbon on major Canadian Prairie crop yields with predictive soil mapping. Agronomy Journal. https://doi.org/10.1002/agj2.21704