| 03 Sep 2025
Soil Health Management Drives Soil Organic Matter More Than Edaphic Properties Across Working Organic Farms
Abstract. Rebuilding soil organic carbon (SOC) on working lands plays a pivotal role in mitigating climate change and improving soil function, yet its accumulation is constrained by both management decisions and inherent soil properties. Scientists and farm advisors recommend that farmers plant cover crops, reduce tillage, and add organic amendments to increase SOC, yet the effectiveness of practices intended to improve soil health may be limited by underlying edaphic controls such as mineralogy, texture, and pH. Given that SOC consists of two distinct fractions—particulate organic matter (POM) and mineral-associated organic matter (MAOM)—which differ in their stability and response to management, a critical question emerges: How much do inherent soil properties limit the effectiveness of recommended soil health practices in increasing SOC? Despite extensive research in controlled field settings, real-world farming contexts remain less understood, limiting our ability to predict SOC gains across diverse soil conditions. Here, we evaluate how in-season and recent (<5 yr) implementation of soil health management systems on working farms affects SOC fractions and stocks across 28 organic fields growing leafy greens in the Central Coast of California. We find that continuous living cover (e.g., through cover cropping) increases three of our measured carbon pools – free POM, MAOM, and surface soil total carbon stocks – while reduced disturbance (i.e., less tillage) increases two – free POM and MAOM. Crop diversity enhances both free and occluded POM fractions. Surprisingly, organic matter amendments do not show any relationship with any of the measured carbon pools. On average, management variables explain 3.7 times more variance than edaphic variables across carbon fractions, whereas, for carbon stocks, the opposite is true: edaphic variables explain ~2.1 times the variance compared to management. Our findings highlight that soil health practices, and in particular continuous cover, can significantly increase soil carbon levels, including both particulate and mineral-associated organic matter fractions, across diverse soil conditions.
Summary Notes of what this paper is about: The authors explore the impact that management may have on soil C pools and overall surface soil C stocks. To do this they incorporated as much variability from four classes of soil management practices across 28 farms into a soil health management variable that they were then able to incorporate into a mixed effects model. They focused on assessing how much variability their model could explain rather than the predictive capacity their model could develop. To this end, I think the authors did a very good job introducing us to the pros and cons of their study. My focus is on trying to ensure that the methodology and model development was clear enough for others to use these same methods to recreate this study in different regions with different crops and soil health management practices. Therefore, my overall advice is for authors to make it a little clearer how they calculated % C pools and how this contrasts from C prop values and mass values. I would suggest putting together a small table like that used in stable isotope studies where folks use calculations to estimate different C pools (Canarini, A., Dijkstra, F.A., 2015. Dry-rewetting cycles regulate wheat carbon rhizodeposition, stabilization and nitrogen cycling. Soil Biology and Biochemistry 81, 195–203). It would just help make the paper much more accessible. I think this kind of study would be insightful to do in other regions with different soil orders, crops, and agricultural practices. Great job! Here are some minor edits I recommend handling before publication:
I’m assuming you looked at surface soils and I think it would be good to just emphasize or connect directly that this reflects your lettuce root depths which would be the zone of influence you wanted to assess, hence you didn’t go deeper. You provide the details but don’t make a direct connection with the writing. Note, it’s not clear to me if this also encompasses the depth of your cover crop plants, it may be shallow enough for catching lettuce rooting impacts but perhaps not for your cover crops where applicable. On this note, you don’t explain what you mean by functional types of crops, define this.
Line 60: You mention MAOM being important for nutrient provisioning but don’t support this with literature, the acknowledgment that there is nutrient provisioning from MAOM is a somewhat new development. Example literature:
Liptzin, D. & Silver, W. L. Effects of carbon additions on iron reduction and phosphorus availability in a humid tropical forest soil. Soil Biol. Biochem. 41, 1696–1702 (2009).
Spohn, M. Preferential adsorption of nitrogen- and phosphorus containing organic compounds to minerals in soils: a review. Soil Biol. Biochem. 194, 109428 (2024).
Similarly, you describe POM fraction C/N ratios but don’t reflect on MAOM ratios – you should include this if you want to bring this in as a comment about POM C/N ratios as it feels like you are trying to develop a comparison. Jilling et al 2018 paper “Minerals in the rhizosphere: overlooked mediators of soil nitrogen availability to plants and microbes” is a good reference for this.
Is the mass of the fPom and oPOM represented in these equations for the mineral+carbon+soil content combined or is this just C mass?
Line 170: You air dried your soils to determine the size of various Fe pools but air drying can lead to irreversible Fe mineralogical changes. In some cases oxidation can lead to Fe(III) formation that then binds to organic matter or changes mineral crystalline structure changing pool make up in your soils. You don’t make a case for different pools in this paper and just refer to total Fe so there is no edit necessary, I just wanted to make sure this was also something that the writers are aware of. For mineralogical assessment of these poos, it may be best to freeze dry. Your sites may have already been pretty dry to begin with so this may not have imposed as much an issue but given that temporally you may have some moisture differences it could have (hard to tell without knowing moisture temporal dynamics from samplings). Kaiser, M., Kleber, M., Berhe, A.A., 2015. How air-drying and rewetting modify soil organic matter characteristics: An assessment to improve data interpretation and inference. Soil Biology and Biochemistry 80, 324–340. Freeze drying has its own issues, but good to just know about.
On Table 1 for Organic Amendments (Total Carbon Input) There seems to be an error, because the Minimum value (4489) is higher than the Max value (2991) – same for the Reduced Disturbance categories, all min values are higher than the max value listed (maybe these were subtracted from 1 so then the input is negatives as described earlier?).
Apologies if this was missed in my review, but I don’t believe your time variable was integrated into your analysis? You mention sampling three times in the year and getting results for each of those sampling times but it isn’t clear in your current manuscript what you did with this resolution of data you collected. Did this add to the analysis or provide any contrast worth discussing? I would assume it did but hard to tell from this current version of the manuscript.
You refer to looking at %C in MAOM and noting that this is not the same as the total MAOM C stock. However, it’s not clear from your description or written equations what the difference is, translation from quantitative term to how you derive this is not clear. I would add a sentence or small example equation to help readers follow. It might behoove the writers to add a table in the supplement where % C in X pool, absolute C in X pool, pool Y stocks are all defined and have an example equation provided for readers and anyone who wants to implement a similar division of C pools in soils for their studies to implement the same calculations. Similarly, I believe when you wrote mean values these are from the entire data set collected whereas C accounting is for C stock of soil (all pools included). But this may need some additional clarification.
I don’t think it’s clear from the description that PC1 and MAOM2 were included to be an edaphic variable grouping in Figure 3 description. Please rewrite it to be clearer.
Line 365: You introduce the term individual practice variables in line 365 as the variables that were better at explaining the variance in the data vs management and edaphic variables but you don’t ever define what individual practice variables are. This makes the text confusing since it seems that here are separate from management variables. Please define these earlier on in the text and make sure your terms referring to subjects involved in the model are consistent.
Line 395 You mention how your results suggest iron and crop diversity as important descriptors for variability in your data and lead to higher C predictions (although you said in your paper that you aren’t really focusing on prediction capabilities for this model?). But you don’t really incorporate this into your discussion of possible mechanisms at play in your system. I think you should include a line that points out at root redox processes that may be allowing iron mineralogy to transform and promote the sticking of aggregates to form oPOM or why these two variables may be contributing together or separately to oPOM increase. Following paper isn’t really looking at oPOM but MAOM does contribute to oPOM formation…so maybe a note for earlier that besides pore space there are other edaphic factors that can contribute to oPOM formation and stability
Yu, G., Xiao, J., Hu, S., Polizzotto, M.L., Zhao, F., McGrath, S.P., Li, H., Ran, W., Shen, Q., 2017. Mineral Availability as a Key Regulator of Soil Carbon Storage. Environmental Science and Technology 51, 4960–4969.
Again, this is kind of hard without knowing what cover crops were used or in what combination. I think the actual identities of the cover crops may make a huge difference in their contribution C and nutrient cycling. There is agricultural literature pointing at these differences in litter/biomass decomposition that could potentially contribute to some of the processes leading to oPOM formation. This second note doesn’t need to be added to the text, but just a note for the authors.
Smychkovich, A., Glaze-Corcoran, S., Keiser, A., Hashemi, M., 2024. Assessing the Root and Shoot Composition, Decomposition, Carbon Contribution and Nitrogen Mineralization Trends of Single Species and Mixed Cover Crops.
Line 435, the sentence structure here makes this difficult to parse, “Future on farm work could also account for interactions between practices and between practices and edaphic conditions. While we account for all practices simultaneously in our models, it is possible that, for example, simultaneous increases in tillage and continuous cover might have an especially positive impact on carbon fractions.”
Also there are some studies on accounting for interactions between soil health/intervention practices in grasslands, maybe include Shangshi paper on this?
Liu, S., Ward, S.E., Wilby, A., Manning, P., Gong, M., Davies, J., Killick, R., Quinton, J.N., Bardgett, R.D., 2025. Multiple targeted grassland restoration interventions enhance ecosystem service multifunctionality. Nat Commun 16, 3971.
Isn’t Fe a micronutrient included in Soil Health Tests for metals? It’s just that the kind of mineralogy that is tied to the presence of Fe and its ability to stabilize C isn’t a measurement, that in itself is too labor intensive of a measurement to be something to include in regular soil health tests at soil health labs.
Line 465, pyrophosphate capitalized when it shouldn’t be
Line 470, edaphic factors like texture (moisture/pore space) and pH may change dynamically over growing seasons due to management to be able to support plant growth. I’d be very curious to know if you get the same trends if you use soils collected from different periods in the growing season and with regards to California rain seasons. (This would be very interesting to also see in regions that experience full seasonality as well).
Line 490 what does directionality mean here?
This leads to the question of what function or ecosystem function does each pool have in these agricultural landscapes because I doubt the C in the MAOM pool is actually holding on to “old C” but its ability to selectively protect certain C compounds and release this in particular conditions may be tied to a function.
What mechanisms generate formation of these different pools, why is it that oPOM and MAOM seem to have formed with strong influence from total Fe variables. This may again tie to the formation of “sticky” Fe oxides that help compact and form aggregates.
Root exudates and enzymes contribute to MAOM formation however Hui Li’s paper on simple metabolites does not show this, this paper is focused on destabilization/disruption, which I don’t think this is an appropriate use of the citation here. I would refer to some of Sokol’s papers for references on MAOM formation from root products. Also: Liang, G., Stark, J., Waring, B.G., 2023. Mineral reactivity determines root effects on soil organic carbon. Nat Commun 14, 4962.
For evidence on root exudate contribution to MAOM formation.
Other references that may be cuing in to the dynamics you suggest about enzyme contribution to MAOM (though in most cases its via fungal enzyme degradation of DOM to then be MAOM)
Klink, S., Keller, A.B., Wild, A.J., Baumert, V.L., Gube, M., Lehndorff, E., Meyer, N., Mueller, C.W., Phillips, R.P., Pausch, J., 2022. Stable isotopes reveal that fungal residues contribute more to mineral-associated organic matter pools than plant residues. Soil Biology and Biochemistry 168, 108634. https://doi.org/10.1016/j.soilbio.2022.108634
Wang, T., Tian, Z., Bengtson, P., Tunlid, A., Persson, P., 2017. Mineral surface-reactive metabolites secreted during fungal decomposition contribute to the formation of soil organic matter. Environmental Microbiology 19, 5117–5129. https://doi.org/10.1111/1462-2920.13990