Mineral-associated organic carbon linked to soil health in farm fields

Farmers who want their fields to fight climate change may have a more powerful tool than previously recognized, according to a new study that finds the most persistent form of soil carbon responds directly to soil management practices. Analyzing soil from 196 agricultural fields on 77 farms across Vermont, researchers led by Erin D. Lane of the USDA Forest Service and Dartmouth College report that the fraction of soil carbon bound to minerals—a form expected to last decades to centuries—was significantly higher in hay and pasture fields than in annually tilled corn and vegetable fields, and that it tracked closely with standard soil health scores.

The findings, published in the journal Biogeochemistry, challenge a lingering assumption in soil science: that management can meaningfully influence the fast-cycling carbon in soil but has little effect on the slow, stable pool. If that assumption had held, the climate benefits of improved farming would have been modest. Instead, the Vermont data suggest that when farmers build soil health, they build durable carbon too.

Soil is the largest active terrestrial carbon reservoir, holding roughly 70 to 80 percent of the carbon stored on land. How much of that carbon stays put depends largely on which of two physical pools it occupies. Particulate organic carbon (POC) consists of partially decomposed plant fragments larger than 53 micrometers—recognizable bits of root and leaf litter that microbes can consume relatively easily. Mineral-associated organic carbon (MAOC), by contrast, is smaller than 53 micrometers and consists of organic compounds chemically bound to clay and silt particles, including the remains of microbes themselves. Because microbes cannot easily reach carbon locked onto mineral surfaces, MAOC persists far longer in soil, making it the prize for any climate mitigation strategy that depends on keeping carbon out of the atmosphere.

To disentangle what controls each pool, the team collected soil samples following the Comprehensive Assessment of Soil Health protocol: ten subsamples of the top 15 centimeters taken in a zigzag pattern across each field, composited and split between Cornell University’s Soil Health Lab and Dartmouth’s Hicks Pries Lab. Fields represented silage corn, hay, pasture, wheat, and mixed vegetable operations, spanning Vermont’s range of soil textures, climates, and management styles. Clay content across sites ranged from about 6 to 54 percent, soil carbon concentrations from 1.45 to 8.68 percent, and mean annual temperature from 4.56 to 8.56 degrees Celsius.

In the laboratory, the researchers performed size fractionation. After drying samples at 90 degrees Celsius and dispersing aggregates with a 0.5 percent sodium hexametaphosphate solution shaken for 24 hours, they wet-sieved the soil through a 53-micrometer mesh. Material retained on the sieve constituted POC; material passing through was MAOC. Each fraction was then dried, ground, and weighed into tin capsules for carbon and nitrogen analysis on an elemental analyzer. Where high pH hinted at carbonate contamination, the team verified inorganic carbon with a hydrochloric acid fizz test and removed it, finding negligible effects on the results.

The numbers told a clear story about management. Hay and pasture fields contained, on average, 69 percent more POC and 30 percent more MAOC than corn silage and vegetable fields. Carbon stocks to 30 centimeters were 17 percent greater in hay fields than in corn fields and 32 percent greater than in vegetable fields. The management history the team could obtain helped explain why: among fields with reported tillage data, 95 percent of pastures and 59 percent of hay fields were in no-till, versus just 7.5 percent of corn fields and 4.5 percent of vegetable fields. Perennial fields also boasted 83 percent greater aggregate stability—resistance of soil clumps to breakdown—which the statistical models identified as a strong positive predictor of both carbon fractions.

Aggregate stability matters mechanistically. Stable macroaggregates, larger than 250 micrometers, physically shelter organic matter from decomposers. Perennial grasses and legumes in hay and pasture contribute abundant root biomass and exudates; belowground inputs from roots outnumber aboveground inputs by roughly eight to one in agricultural systems, and roots feed the microbial activity that both breaks down POC and generates the microbial necromass that ultimately becomes MAOC. Manure inputs, common on livestock farms, likely accelerate this conversion by fueling decomposers.

Climate and texture left their fingerprints as well, in ways that depended on one another. POC concentrations rose with precipitation at cooler sites but flattened at warmer ones, consistent with moisture and temperature jointly governing the balance between plant input and microbial decomposition. More strikingly, MAOC responded to a three-way interaction among clay content, mean annual temperature, and mean annual precipitation. In clay-rich soils, MAOC increased with warming, especially where rainfall was abundant—warmer, wetter conditions boost plant productivity and microbial processing, delivering more dissolved organic carbon onto abundant mineral surfaces. In low-clay soils the pattern reversed or vanished: with fewer sorptive surfaces available, intensified decomposition under warm, wet conditions led to carbon loss through leaching and mineralization rather than stabilization.

Crucially for the carbon sequestration debate, the team found that Vermont’s farm soils remain well below their mineral saturation point. Comparing MAOC concentrations against the relationship between clay-plus-silt content and MAOC established by a recent global analysis, the measured values fell far beneath the saturation threshold. That means these soils retain genuine capacity to lock away more persistent carbon—capacity that management can exploit. Perennial crops, the data suggest, are among the most effective ways to do so, simultaneously maximizing both POC and MAOC.

Perhaps the most consequential result is the link to soil health scoring. The Cornell framework assigns each field a composite score integrating biological, chemical, and physical metrics such as active carbon, aggregate stability, pH, and organic matter content. The researchers found that every one-unit increase in the soil health score corresponded to a 3.15 percent increase in POC concentration and a 2.25 percent increase in MAOC concentration. Carbon stocks to 30 centimeters rose by roughly 0.25 kilograms of carbon per square meter per score unit, and the relationship was tight, explaining 60 percent of the variance in stocks. A commonly used quick indicator, permanganate-oxidizable “active” carbon, also correlated positively with both fractions, though it proved the least sensitive of the carbon measures to differences among crop types.

One nuance deserves attention: while absolute MAOC rose with soil health scores, the proportion of total carbon in mineral-associated form declined slightly, because POC accumulated even faster. Corn fields actually showed a higher MAOC proportion than pastures—not because they held more stable carbon, but because tillage and low residue inputs had stripped away the vulnerable POC, leaving the remainder proportionally enriched. A high MAOC proportion, in other words, can signal degradation rather than health.

The study’s observational design means management and environment cannot be fully separated—farmers choose crops partly based on the land they have. But the authors note that the one consistently significant difference among crop types, aggregate stability, is a soil property highly responsive to management. Soil pH, often suspected of controlling carbon dynamics, played no significant role here, likely because lime applications kept fields within a narrow and favorable range.

The implications reach beyond Vermont’s dairy belt. Meta-analyses across Europe and the Americas have found cover cropping increases POC by about 15 percent and MAOC by 5.6 percent, and studies in semi-arid systems show similar responsiveness to perennial vegetation. If farmers, agricultural extension programs, and carbon-credit markets can use routine soil health tests as a proxy for durable carbon gains, the pathway to financing regenerative practices becomes considerably simpler. As the authors conclude, building soil health and building long-term soil carbon are, it turns out, far more aligned than they are in conflict.