Soil organic matter (SOM) encompasses a diverse array of biomolecules, spatially and temporally distributed across multiple biophysical gradients (Lehmann & Kleber, 2015). Soil particulate organic matter (POM) and mineral-associated organic matter (MAOM) are two physically distinct pools with notable differences in sources, chemical composition, and residence time (Angst et al., 2021). POM, predominantly originating from partially decomposed plant debris, typically possesses a shorter residence time and is prone to decomposition, especially when not embedded within soil aggregates (von Lützow et al., 2007). Conversely, MAOM, characterized by its tight association with soil minerals or encapsulation within microaggregates (<50 μm), is thought to be more persistent in soils, with residence times ranging from decades to centuries (Sokol et al., 2019). A clear understanding of the plant and microbial sources of POM and MAOM formation is critical to informing SOM dynamics and stability. However, prevailing methodologies, including microbial biomarker analysis, molecular fingerprinting, and mathematical modeling, encounter critical limitations. These limitations are (1) the erroneous assumption of uniform biomarker distribution between biomass and soil; (2) the oversimplified attribution of compound classes to specific biological sources; and (3) unreliable conversion metrics and indirect SOM quantification methods (Whalen et al., 2022). Furthermore, no consensus has been reached regarding the relative contributions of microbial and plant sources to SOM (Angst et al., 2021).

The recent work of Chang et al. (2024) provides another innovative angle to advance the understanding of MAOM sources through their application of a two-pool mixing model, grounded in the principles of ecological stoichiometry. Based on this model, Chang et al. (2024) quantified plant and microbial contributions to MAOM across diverse ecosystems, including forests, grasslands, and croplands, based on an extensive dataset of 288 samples. Contrary to the current understanding, their results suggest more balanced contributions from plant and microbial sources to MAOM, with microbial inputs accounting for 34%–47% and plant residues for 53%–66%. This discovery fundamentally challenges the conventional understanding of microbial-dominated MAOM formation, highlighting the potential of a two-pool mixing model as an economically viable supplement to microbial biomarker and molecular fingerprinting approach in identifying the sources of MAOM.

However, we have two concerns regarding their framework. First, contrary to the methodology employed by Chang et al. (2024) that assumes POM as the sole plant source for MAOM, dissolved organic carbon from partially decomposed plant litters can also be adsorbed by minerals to form MAOM. For instance, Sokol et al. (2019) found that dissolved organic carbon can be a major carbon source in MAOM, especially in high-leaching forest ecosystems, where dissolved organic matter can contribute up to 89% of the carbon in MAOM. Second, Chang et al. (2024) adopted a constant C:N ratio over time, yet C:N ratios for plant and microbial sources vary across temporal scales. Indeed, temporal variations in litter C:N ratios have been observed in various ecosystems (Liu, Fang, et al., 2023; Liu, Qiu, et al., 2023; McGroddy et al., 2004), reflecting adaptations to nutrient requirements and environmental changes during different growth phases. Similarly, soil microorganisms also exhibit shifts in composition in response to environmental and phenological changes, leading to temporal variations in microbial C:N ratios (Liu, Qiu, et al., 2023; Xu et al., 2013). Therefore, we speculate that the unrealistic estimation of plant or microbial contributions to MAOM (i.e., >100%) in Chang et al. (2024) could likely be attributed to these concerns.

Here, in addition to considering plant litter as a plant source for MAOM, we emphasize the importance of incorporating dynamic C:N ratios for both plant and microbial sources in determining the origins of MAOM. Nevertheless, the concurrent measurement of C:N ratios in plant litter and microbial communities over prolonged periods remains scarce, presenting significant challenges to the practical application of our novel framework. To address this challenge, we advocate for large-scale spatial sampling as an effective alternative to temporal sampling, covering multiple temporal stages of biological growth. To overcome limitations imposed by temporal variations and enhance global applicability, we highlight the necessity of establishing the variation in C:N ratios across different ecosystems globally.

Applying ecological stoichiometry to trace SOM sources offers significant theoretical, practical, and modeling advancements. First, this method integrates ecological stoichiometry principles into tracing SOM sources, providing a novel approach based on elemental balance theory. Second, this method enables a comprehensive global assessment of the sources of SOM, utilizing extensive global databases that document the biogenic elemental composition (such as carbon and nitrogen) of plants and soil microorganisms. Lastly, the formulas derived from this framework hold considerable promise for inclusion in data-driven models, potentially enhancing the accuracy of model projections related to soil carbon and nitrogen cycling.

Acknowledging the inherent limitations of this stoichiometric approach is crucial for refining its accuracy and applicability. First, a primary constraint lies in the method's overlook of the distinct C:N ratios exhibited by living versus dead soil microorganisms. The ecological stoichiometry of these organisms certainly shifts upon death—for instance, soil microorganisms in nitrogen-poor environments might enhance nitrogen sequestration during metabolic processes, and nitrogen from dead microorganisms tends to be more readily assimilated by surviving organisms. This dynamic can lead to dead microbial matter having a higher C:N ratio than its living counterpart (Liu, Qiu, et al., 2023). It should be noted that prevalent methods of measuring microbial carbon and nitrogen mainly rely on chloroform fumigation extraction, and primarily measure elements released by living microorganisms rather than microbial residues, thereby introducing uncertainties when using these measurements to infer the C:N ratios of microbial residues. This limitation underscores the importance of developing refined methodologies for accurately determining the C:N ratios of external sources.

Second, our methodology primarily targets identifying SOM sources within the topsoil layer, encountering significant uncertainties when applied to subsoil SOM tracing. Current datasets largely focus on senescent leaves and branches, inadequately accounting for root contributions to SOM formation in subsoil layers. The method also does not consider variations in soil microbial community composition and microbial residue stoichiometry across different soil depths. This oversight restricts the method's effectiveness in precisely identifying SOM sources below the topsoil layer, pointing to a critical need for innovative approaches that can evaluate SOM sources throughout the entire soil profile.

In summary, the method introduced by Chang et al. (2024) for quantifying plant and microbial contributions to MAOM represents a significant advancement in understanding SOM sources. Building on this groundwork, we propose a more comprehensive Bayesian framework that incorporates nuanced variations in ecological stoichiometry, aiming to effectively improve the accuracy of global SOM source assessments. Future research endeavors should identify the stoichiometric differences between living and deceased microbial sources, as well as delineate the dynamics of ecological stoichiometric ratios among plant sources, microbial sources, and SOM across different soil layers. Such investigations will not only advance the understanding of SOM sources and the model prediction of SOM dynamics under future climate scenarios but also can help optimize ecosystem management toward climate mitigation and sustainable agriculture.

土壤有机质 (SOM) 包含多种生物分子，在空间和时间上分布在多个生物物理梯度上（Lehmann & Kleber， 2015 ）。土壤颗粒有机质（POM）和矿物相关有机质（MAOM）是两个物理上不同的池，在来源、化学成分和停留时间方面存在显着差异（Angst等人， 2021 ）。 POM 主要源自部分分解的植物残骸，通常具有较短的停留时间并且易于分解，特别是当未嵌入土壤团聚体时（von Lützow 等， 2007 ）。相反，MAOM 的特点是与土壤矿​​物质紧密结合或封装在微团聚体中 (<50 id=3>2019)。清楚了解 POM 和 MAOM 形成的植物和微生物来源对于了解 SOM 动态和稳定性至关重要。然而，流行的方法，包括微生物生物标志物分析、分子指纹识别和数学建模，遇到了严重的局限性。这些局限性是（1）生物质和土壤之间生物标志物分布均匀的错误假设； (2) 将化合物类别过于简单化地归因于特定的生物来源； (3) 不可靠的转换指标和间接 SOM 量化方法（Whalen 等人， 2022 ）。此外，关于微生物和植物来源对 SOM 的相对贡献尚未达成共识（Angst 等人， 2021 ）。

Chang 等人的最新工作。 ( 2024 ) 通过应用基于生态化学计量原理的两池混合模型，提供了另一个创新角度来促进对 MAOM 来源的理解。基于该模型，Chang 等人。 ( 2024 ) 基于 288 个样本的广泛数据集，量化了包括森林、草原和农田在内的不同生态系统中植物和微生物对 MAOM 的贡献。与目前的认识相反，他们的结果表明植物和微生物来源对 MAOM 的贡献更加平衡，微生物输入占 34%–47%，植物残留物占 53%–66%。这一发现从根本上挑战了对微生物主导的 MAOM 形成的传统理解，凸显了两池混合模型作为微生物生物标志物和分子指纹方法在识别 MAOM 来源方面的经济上可行的补充的潜力。

然而，我们对其框架有两个担忧。首先，与 Chang 等人采用的方法相反。 ( 2024 )假设POM是MAOM​​的唯一植物来源，部分分解的植物凋落物中溶解的有机碳也可以被矿物质吸附形成MAOM。例如，索科尔等人。 ( 2019 )发现溶解有机碳可以成为MAOM的主要碳源，特别是在高淋滤森林生态系统中，溶解有机碳可贡献高达MAOM碳的89%。其次，张等人。 （ 2024 ）随着时间的推移采用了恒定的 C:N 比率，但植物和微生物来源的 C:N 比率在不同时间尺度上有所不同。事实上，在不同的生态系统中都观察到了凋落物碳氮比的时间变化（Liu, Fang, et al., 2023 ；Liu, Qiu, et al., 2023 ；McGroddy et al., 2004 ），反映了对养分需求的适应以及不同生长阶段的环境变化。同样，土壤微生物也会因环境和物候变化而表现出成分变化，导致微生物 C:N 比率随时间变化（Liu, Qiu, et al., 2023 ；Xu et al., 2013 ）。因此，我们推测 Chang 等人对植物或微生物对 MAOM 的贡献（即 >100%）的不切实际的估计。 （ 2024 ）可能归因于这些担忧。

在这里，除了将植物凋落物视为 MAOM 的植物来源外，我们还强调在确定 MAOM 起源时结合植物和微生物来源的动态 C:N 比率的重要性。然而，长期同时测量植物凋落物和微生物群落中的碳氮比仍然很少，这对我们的新框架的实际应用提出了重大挑战。为了应对这一挑战，我们主张大规模空间采样作为时间采样的有效替代方案，涵盖生物生长的多个时间阶段。为了克服时间变化带来的限制并增强全球适用性，我们强调有必要确定全球不同生态系统之间 C:N 比率的变化。

应用生态化学计量学追踪 SOM 来源提供了重大的理论、实践和建模进步。首先，该方法将生态化学计量原理融入追踪SOM来源，提供了一种基于元素平衡理论的新方法。其次，该方法利用记录植物和土壤微生物的生物元素组成（如碳和氮）的广泛全球数据库，能够对 SOM 来源进行全面的全球评估。最后，从该框架得出的公式有望纳入数据驱动模型，从而有可能提高与土壤碳和氮循环相关的模型预测的准确性。

承认这种化学计量方法的固有局限性对于提高其准确性和适用性至关重要。首先，主要的限制在于该方法忽略了活土壤微生物与死土壤微生物所表现出的不同的 C:N 比率。这些生物体的生态化学计量在死亡时肯定会发生变化，例如，贫氮环境中的土壤微生物可能会在代谢过程中增强氮固存，而死亡微生物中的氮往往更容易被幸存的生物体同化。这种动态可能导致死亡的微生物物质比其活体微生物物质具有更高的 C:N 比率（Liu, Qiu, et al., 2023 ）。值得注意的是，测量微生物碳和氮的普遍方法主要依靠氯仿熏蒸提取，并且主要测量活体微生物释放的元素而不是微生物残留物，因此在使用这些测量值推断微生物残留物的C：N比时引入了不确定性。这一限制强调了开发精确方法来准确确定外部来源的 C:N 比率的重要性。

其次，我们的方法主要针对识别表土层内的 SOM 来源，在应用于底土 SOM 追踪时遇到很大的不确定性。当前的数据集主要集中在衰老的叶子和树枝上，没有充分考虑根系对底土层中 SOM 形成的贡献。该方法也没有考虑不同土壤深度的土壤微生物群落组成和微生物残留化学计量的变化。这种疏忽限制了该方法在精确识别表土层以下 SOM 源方面的有效性，表明迫切需要能够评估整个土壤剖面中 SOM 源的创新方法。

总之，Chang 等人介绍的方法。 ( 2024 ) 量化植物和微生物对 MAOM 的贡献代表了理解 SOM 来源的重大进步。在此基础上，我们提出了一个更全面的贝叶斯框架，其中纳入了生态化学计量的细微差别，旨在有效提高全球 SOM 源评估的准确性。未来的研究工作应确定活微生物源和死亡微生物源之间的化学计量差异，并描绘不同土层植物源、微生物源和 SOM 之间生态化学计量比的动态。此类研究不仅将增进对 SOM 来源的理解以及未来气候情景下 SOM 动态的模型预测，而且有助于优化生态系统管理，以实现气候缓解和可持续农业。