| 12 Dec 2025
Magnetic separation reveals overestimation of soil organic matter due to undecomposed particulate residues
Abstract. Soil organic matter (SOM) is a complex mixture of organic compounds derived from the decomposition of plant and animal residues. Only after undergoing microbial transformation and forming stable associations with minerals can it be considered “true” SOM, rather than a simple mechanical accumulation of carbon-containing substances. According to current understanding, particulate organic matter (POM) comprises both undecomposed and partially decomposed organic residues. Of these, the undecomposed fraction does not qualify as SOM in the strict sense. However, conventional analytical methods cannot fully distinguish fine particulate residues from soil matrices, leading to an overestimation of POM-derived carbon content. The extent and persistence of this “false increase” in SOM due to “disguised” POM remains poorly understood. In this study, straw and biochar were magnetized via chemical co-precipitation and applied to soils. The incompletely decomposed magnetized residues in the soil were separated using magnetic separation at various time points, enabling more accurate tracking of SOM dynamics. Five treatments were established: blank control (CK), untreated straw (CS), untreated biochar with carbon input equivalent to untreated straw (Bc), magnetized straw with carbon input equivalent to untreated straw (MCS), and magnetized biochar with carbon input equal to untreated straw (MBc). The results showed that after the application of organic materials into the soil, the recovery rate of magnetized straw residues declined continuously, reaching 54.55 % after 360 d, whereas biochar remained largely stable at 92.48 %. In CS and Bc treatments, the organic carbon content of POM fractions and their proportion in the total SOM were consistently higher than in CK, particularly during early incubation. However, this was attributable to overestimation from incompletely decomposed residues. In contrast, MCS-D and MBc-D treatments (after magnetic residue removal) showed minimal deviation from CK, confirming the contribution of incomplete decomposition to SOM overestimation. On day 30, the apparent increase in the particulate organic carbon (POC) content reached 63.48 % for CS and 58.99 % for Bc. Over time, the overestimation in the CS treatment declined to 15.34 % by day 360, whereas the overestimation in the Bc treatment remained largely unchanged, with a 53.71 % increase persisting. These findings highlight the potential for SOM overestimation when POM fractions are fully included without accounting for undecomposed inputs, particularly the long-term persistence of recalcitrant organic materials, which may introduce systematic bias in global SOM quantification.
Xia et al. present an interesting study in which straw and biochar are magnetized prior to an incubation experiment, enabling magnetic separation to remove incompletely decomposed amendment residues. This allowed the authors to track the persistence of amendment-derived particulate material over one year and to evaluate the extent to which observed increases in the POM fraction reflect persistent amendment fragments versus processed organic matter.
The study fits well within SOIL’s scope and presents a potentially novel methodological approach to addressing a common interpretation issue in SOM fractionation studies. However, I recommend major revisions, primarily because (1) the conceptual framing regarding what “counts” as SOM is not fully aligned with current usage in soil science, (2) several methodological choices require clearer justification and/or more explicit discussion of limitations, and (3) the conclusions are broader than the evidence supports. Minor issues include occasional unclear wording and a few instances where the citation support appears weak or poorly targeted.
General comments
My main concern is that the manuscript is asserting a definition of SOM that is not consistent with how it is commonly defined and used in current soil science. In particular, the text implies that only mineral-associated OM represents “true SOM,” and that undecomposed POM does not qualify as SOM. This framing risks conflating two different issues: whether carbon is present in the soil system (which can legitimately increase after organic additions), versus whether that carbon is stabilized/persistent (e.g., mineral-associated and/or physically protected) and therefore likely to contribute to longer-term SOC storage.
I suggest reframing the key message away from “overestimation of SOM/SOC” and toward the also important point that increases in operational POM fractions following amendment addition can be over-interpreted as evidence of stabilization or processed SOM formation, when they may instead reflect persistent, untransformed amendment residues. The manuscript would benefit from grounding the introduction more explicitly in the POM–MAOM conceptual framework and related literature, perhaps including a simple conceptual diagram distinguishing amendment residues vs processed OM and where each falls within SOM fractions.
A further general issue is the statistical framework. Given the treatment x time design with destructive sampling, the manuscript would benefit from a model-based approach (e.g., two-way ANOVA/linear models with treatment, time, and interaction, followed by appropriately corrected post-hoc comparisons) rather than relying primarily on Duncan’s multiple range test.
The main methodological contribution (using magnetic separation to quantify residue contribution to operational POM/POC) is promising and likely of interest to the community. However, the concluding statements should be revised to avoid implying that POM is not part of SOM and to substantially moderate the leap from this single-soil incubation study to claims of “systematic bias in global SOM content assessments.”
Specific comments
Line 47-48: What exactly is meant by “with the greatest binding to minerals”?
Line 49-52: What is meant by “the core of SOM”? This sentence is unclear to me, and I am not sure that Feng et al. 2025 should be the only citation included, given all of the claims listed.
Line 107-109: How were these properties measured?
Line 174: Should read “soil organic carbon (SOC)” rather than “soil organic matter”
Line 175-178: Please clarify carbon quantification methods. Was the dichromate oxidation method used only for the original soils, or also for fractions? Why was elemental analysis not used consistently for SOC/OC across soil/fraction samples?
Line 181-182: Eqs. (1)–(2) are currently confusing due to unit/definition issues. MP is defined as a percentage mass proportion of the POM fraction, but Eq. (1) (POC = MP × OCP) only yields correct units if MP is treated as a mass fraction (MP/100) rather than a percent. Please clarify MP (fraction vs %) and adjust Eq. (1) accordingly. Additionally, Eq. (2) labeled “POM (%)” is actually POC/SOC × 100, i.e., the contribution of POM-C to total SOC, not the proportion of POM mass; the variable name should be changed to avoid misinterpretation.
Line 199: Please clarify the statistical framework. As written, only Duncan’s multiple range test is reported, but the manuscript does not state the test/model from which the post-hoc comparisons derive (e.g., one-way ANOVA at each time point vs a factorial treatment × time model). Given the incubation design, it is important to test treatment, time, and their interaction, and to specify the corresponding model structure and assumptions checks. I recommend a two-way ANOVA/linear model (treatment, time, treatment × time) with appropriately corrected post-hoc comparisons, or explicit justification if a different approach was used.
Lines 254-260: The mechanistic language around “molecular conjugation,” “electronic environment,” and “enhanced aromatic structural features” is overly specific for the bulk elemental/Van Krevelen data in Fig. 2; please rephrase these interpretations more cautiously (e.g., shifts in H/C and O/C consistent with oxidative processing and changes in aliphaticity/condensation), and avoid “humification” unless operationally defined and supported with appropriate chemical evidence.
Line 290: “Artificial elevations” is misleading. The higher POM mass proportion reflects persistent amendment fragments captured in the operational >53 µm fraction; please rephrase as something like “residue-driven increase.”
Line 301: Same comment as above regarding “false increase”
Section 4.1: Again please tone down the highly specific mechanistic language inferred from Fig. 2 (e.g., “molecular conjugation,” “electronic environment,” “humification”) and avoid attributing straw–biochar differences primarily to C/N; biochar persistence is strongly controlled by feedstock/pyrolysis chemistry. Additionally, the long list of generic biochar benefits reads like background review and should be shortened or linked explicitly to your results.
Section 4.2: The framing of a “false/artificial elevation” and “true organic matter” is misleading. The operational POM fraction can legitimately increase after amendments; the key issue is interpretation (residue-derived POM vs processed/stabilized C), not that SOM/SOC is “overestimated.” Please rephrase accordingly, avoid implying POM is not SOM, and significantly moderate/justify the leap to “overestimation of global SOM content.”