Farm Management Beats Soil Type in Explaining Carbon on Working Organic Farms

On working organic vegetable farms along California’s Central Coast, the way farmers manage their land appears to matter more for soil carbon than the inherent properties of the soil beneath it. A new study of 28 organic fields growing leafy greens in Monterey, Santa Cruz, and San Benito Counties found that management variables explained roughly 3.7 times more of the variation in soil carbon fractions than edaphic variables such as texture, pH, and iron mineralogy. Only for total carbon stock did the balance reverse, with soil properties explaining about 2.1 times more variance than management. The findings, published in the journal SOIL, suggest that practices like keeping living roots in the ground year-round and reducing tillage can build soil carbon across a strikingly wide range of soil conditions, even on commercial farms where those practices are applied unevenly and imperfectly.

The research team, led by Kenzo Esquivel of the University of California, Berkeley, focused on a question that has long complicated soil carbon science: do soil-building practices actually work in the messy, real-world conditions of working farms, or are their effects overwhelmed by the inherent characteristics of each soil? Most previous evidence comes from controlled field experiments, where researchers can isolate the effect of a single practice while holding soil type constant. Those experiments are valuable, but their results can be difficult to extrapolate to real farms, where practices vary in intensity and consistency and where soils differ dramatically from field to field. On-farm studies complement controlled experiments by capturing how diverse practices and soil types interact under actual production conditions.

The study zeroed in on a distinction that is increasingly central to soil carbon science. Soil organic carbon is not a single substance but a collection of fractions with different properties and fates. Particulate organic matter, or POM, consists of partially decomposed plant material that cycles relatively quickly, feeding microbes and decomposers while improving soil structure, aggregate stability, and water infiltration. POM can be further divided into free POM, which sits unprotected in the soil, and occluded POM, which is physically bound inside soil aggregates and thus partially shielded from microbial decay. Mineral-associated organic matter, or MAOM, is different in kind: it consists of simpler carbon molecules that form chemical and physical bonds with mineral surfaces, protecting them from decomposition and allowing them to persist for decades, centuries, or in some cases millennia. POM is generally tied to near-term soil fertility, while MAOM serves as a longer-term carbon reservoir.

To measure these fractions, the researchers sampled soils from 28 organic lettuce fields during the 2020 growing season, collecting composite samples from the top 15 centimeters of soil along transects running through the middle of each field. Using a joint density and size fractionation protocol, they separated the soil into dissolved organic matter, free POM, occluded POM, and MAOM, then measured the carbon content of each fraction by combustion analysis. They also measured bulk density, texture, pH, cation exchange capacity, and a suite of iron phases using sequential chemical extractions, since iron oxides and organo-mineral complexes provide reactive surfaces where carbon can sorb and stabilize. Management information came from in-depth surveys conducted with farmers, covering the previous five years of cover cropping, tillage, crop diversity, and organic amendments, supplemented by satellite imagery analysis that quantified the proportion of each year, from 2015 through 2019, that each field had living vegetation cover.

The clearest signals emerged for particulate carbon. Fields with higher continuous living cover over the previous five years had substantially more free POM: the model predicted that shifting from a site with 23 percent of the year under vegetation to one with 90 percent cover would add about 3.16 megagrams of carbon per hectare in the free POM fraction, with a 95 percent confidence interval of 0.82 to 5.92. Reduced tillage showed a similar pattern, with moving from deep tillage more than once per year to never deep tilling associated with roughly 1.58 megagrams of additional free POM carbon per hectare. Crop diversity was linked to occluded POM, with a model comparison between a low-diversity field with three cash crops and a high-diversity site with six cash crops and nine cover crop species predicting an increase of 1.35 megagrams of carbon per hectare in that aggregate-protected fraction.

The associations with mineral-associated carbon and total carbon stock were weaker and less statistically clear. Continuous living cover and reduced tillage were both positively associated with MAOM carbon concentration, but the relationships fell at or just outside conventional significance thresholds, with p-values between 0.07 and 0.11 depending on model specification. The modeled effects were on the order of 0.6 percentage points of MAOM carbon concentration. For total surface carbon stock, continuous cover showed a borderline association, with the model estimating an increase of about 10.15 megagrams of carbon per hectare in the top 15 centimeters when comparing high- and low-cover sites, though the confidence interval ranged from slightly negative to strongly positive. The authors are careful to frame these as tentative findings rather than confirmed effects, noting that the five-year management window captured in their data may be too short for MAOM accumulation, which meta-analyses suggest becomes more pronounced in experiments lasting a decade or longer.

One result stood out as unexpected: organic amendments, including compost, composted chicken manure, and commercial organic pellet fertilizers, showed no association with any measured carbon pool. This contradicts controlled studies that have documented soil carbon gains following compost and manure application. The researchers offer several possible explanations. Their amendment measure covered only the 2020 sampling season, on the assumption that annual practices are relatively stable, but soil carbon accumulation reflects inputs over much longer timescales, so even a representative annual rate may miss the cumulative amendment history most relevant to detecting a response. A bivariate check also revealed a negative correlation between carbon inputs and POM fractions, consistent with a possible priming effect, in which nutrient-rich amendments stimulate microbial activity and accelerate the decomposition of existing particulate organic matter. The authors also note that amendments did correlate with crop yield in a companion analysis, suggesting these inputs have agronomic effects even if carbon outcomes were not detectable here.

Edaphic factors were far from irrelevant. Iron mineralogy emerged as an informative variable in most models, with poorly crystalline iron and organo-mineral complexes showing clear positive associations with MAOM carbon concentration, consistent with their known role in providing sorption sites for carbon stabilization. Iron phases were also linked to POM accumulation, most clearly for occluded POM, possibly because short-range-order iron minerals promote aggregate stability and physically protect organic matter. Yet despite iron’s consistent presence, management variables explained more variance than edaphic variables for all three carbon fractions. This was somewhat surprising for MAOM, given the established role of soil pH and texture in sorption dynamics. The study sites spanned an unusually wide edaphic range for a single region, with clay content from 5 to 60 percent and sand from 5 to 77 percent, and the fact that management signals held up across that variability is itself informative for how the findings might generalize.

Total carbon stock was the one outcome where soil properties dominated, which the authors attribute to the way silt and clay content increases the surface area available for carbon sorption. They also addressed a potential confound: because reduced tillage might simply redistribute carbon deeper into the profile rather than increase it, they compared bulk carbon at three depth increments across the disturbance gradient. They found no evidence of a surface carbon deficit at more-disturbed sites, and if anything, less-disturbed sites showed significantly higher carbon at 30 to 60 centimeters depth. This pattern is more consistent with reduced tillage supporting whole-profile carbon accumulation than with redistribution of a fixed pool, though the authors emphasize that fraction-level sampling at depth would be needed to confirm this directly. They also stress that their observational design cannot establish causation; the associations are consistent with proposed mechanisms but do not prove them.

The practical implications are considerable. Cover crops remain rare in the study region, with a recent remote sensing analysis finding that only about 6 percent of Central Coast farmland had winter cover crops and roughly 60 percent of farmers left fields fallow through the winter. Nationally, only 5.6 percent of available United States cropland is planted with cover crops. Reduced tillage is similarly uncommon in California vegetable production. This means there is substantial untapped opportunity for building particulate soil carbon through practices that the evidence links most clearly to carbon gains. The findings also carry a caution for policy: California’s Healthy Soils Program allocated more than two-thirds of its grant funding, nearly 56 million dollars in 2021 alone, toward compost and carbon amendments, yet this study found no measurable association between amendments and any carbon pool on these farms. Whether the weaker signals for mineral-associated carbon and total stock strengthen over longer timescales remains a key open question for whether soil health management can contribute to long-term carbon sequestration, and not just near-term soil function, across the diverse soils that farmers actually work.

Subject of Research: Soil health management practices and soil organic carbon fractions on working organic farms

Article Title: Soil health management outweighs edaphic properties in explaining soil organic matter across working organic farms

Article References: Esquivel, K. E., Waterhouse, H., Thompson, J., Karp, D. S., Santos, G., Gil-Santos, Y., Baur, P., Iles, A., & Bowles, T. (2026). Soil health management outweighs edaphic properties in explaining soil organic matter across working organic farms. SOIL, 12(2), 1035-1055. https://doi.org/10.5194/soil-12-1035-2026

Image Credits: AI Generated

DOI: 10.5194/soil-12-1035-2026

Keywords: soil organic carbon, particulate organic matter, mineral-associated organic matter, cover crops, reduced tillage, crop diversity, organic farming, soil health, carbon sequestration, edaphic properties, California Central Coast, iron mineralogy