the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Jul 2026
The Implications of Microbial Functional Diversity on the Response of Soil Organic Matter Decomposition to Nitrogen Fertilization: A Theoretical Model
Abstract. Nitrogen (N) fertilization decreases carbon (C) decomposition, but there remain competing hypotheses underlying this response including reduced belowground C allocation, shifts in microbial diversity, and altered microbial function. Further, current soil organic matter (SOM) decomposition models lack the ability to test these hypothesized mechanisms. To theoretically test these mechanisms, we integrated the Carbon Acquisition Ecological Strategies (CAES) framework into the Carbon, Organisms, Rhizosphere, and Protection in the Soil Environment (CORPSE) model. CAES represents a pathway for integrating functional diversity into microbial explicit soil decomposition models, with three microbial functional groups that specialize in decomposing and utilizing a different SOM pool: primary decomposers (1° decomposers) target complex SOM, secondary decomposers (2° decomposers) target microbial necromass, and passive consumers target simple SOM. We then used the refined model (Functional Group-CORPSE) to theoretically test hypotheses by leveraging long term data from a 30 year, watershed scale N fertilization experiment at the Fernow Experimental Forest in Parsons, WV. We performed modeling experiments to evaluate the microbial community response to varying C and N inputs, and to simulate hypothesized mechanisms of microbial response to N deposition. Varying substrate composition led to distinct microbial communities, reflecting substrate preferences between functional groups. However, only when both the function and abundance of complex C decomposers was reduced could FG-CORPSE capture the SOM dynamics observed at the Fernow. Collectively, our results show that FG-CORPSE enables the simulation of both functional and compositional shifts in microbial communities to facilitate the testing of microbe centric hypotheses for decomposition and C storage under N enrichment.
- Preprint
(1132 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-3566', Anonymous Referee #1, 14 Aug 2026
-
RC2: 'Comment on egusphere-2026-3566', Anonymous Referee #2, 23 Sep 2026
Overall comments
This study debuted a new microbially explicit soil C model that combines the Carbon Acquisition Ecological Strategies (CAES) framework into the Carbon, Organisms, Rhizosphere, and Protection in the Soil Environment (CORPSE) model. The authors then test the model by changing litter substrate inputs and evaluating sensitivity of microbial groups to those changes. They also apply the model for an N deposition experiment, to evaluate whether altering decomposer functions leads to better observation-model alignment. They find their microbial groups are slightly sensitive to litter input chemistry and that the observations and model output align when primary decomposer decomposition rate and turnover are modified. I think this paper has an important and interesting focus and is impressive in its presentation of a new model that is able to reproduce observed responses. However, there are some areas where the text could improve:
- The introduction would benefit form a more nuanced argument for the need for this model – I provide specific comments below.
- The hypothesized changes under N addition only address primary decomposers which feels a bit odd given this work is focused on integrating microbial functional groups into CORPSE. The authors should justify this better.
- The litter chemistry effects are subtle, and this should also be addressed more clearly.
- Adding the experiment where only mineral N input and root inputs were changed to the main text results and figures would make the argument for the need for microbial functional representation much stronger, as these results show that the soil C values are only aligned between observations and model output when microbial functions are changed.
Specific comments
Line 32: This is too non-specific – decomposition of litter or soil? And I am not sure this is consistent enough to state this outright. Same comment applies to line 62 – there are examples of N fertilization reducing SOC (e.g., https://doi.org/10.1016/j.soilbio.2015.10.002) – perhaps frame this as the general response.
Line 33-33: Shifts in microbial diversity likely explain functional shifts, right? I am not sure these should be separate drivers.
Lines 64-66: Mentioning acidification seems important here as it is often highlighted as a mechanism.
Line 68: Are there other examples beyond Eastman et al., 2024 that show first-order models consistently misrepresent this? Notably, Eastman et al., 2024 reports that both the first-order and MIMICS models can represent soil C responses. I think the argument needs to be more nuanced here.
Line 83 and throughout: Minor thing, but it is a bit more intuitive to define the whole acronym and then put the acronym in the parentheses.
Line 86-88: r- and K-strategists differently contribute to POM and MAOM-like pools in MIMICS, so I wouldn’t say those cannot be linked. Perhaps a bigger challenge with MIMICS is the ability to link r- and K-strategists to measured microbial data. Further, Eastman et al., 2024 shows MIMICS can capture these responses so, again, a more nuanced argument could be used here. Further, it might be helpful to briefly explain why POM and MAOM are of interest and how they are different, since that is a key finding at Fernow.
Line 91: I also wouldn’t say CORPSE is really focused on measurable pools – that’s more the focus of MEMS (Zhang et al., 2021) or Millennial (Abramoff et al., 2022). Further I don’t believe CORPSE has been directly tested for its ability to capture N fertilization responses, so it may be able to. Perhaps rather than saying these responses cannot be captured, the authors could focus on the mechanisms that cannot be assessed in current models and how their model provides a novel area for mechanistic testing.
Abramoff, R.Z., Guenet, B., Zhang, H., Georgiou, K., Xu, X., Rossel, R.A.V., Yuan, W., Ciais, P., 2022. Improved global-scale predictions of soil carbon stocks with Millennial Version 2. Soil Biology and Biochemistry 164, 108466.
Zhang, Y., Lavallee, J.M., Robertson, A.D., Even, R., Ogle, S.M., Paustian, K., Cotrufo, M.F., 2021. Simulating measurable ecosystem carbon and nitrogen dynamics with the mechanistically defined MEMS 2.0 model. Biogeosciences 18, 3147-3171.
Lines 108-117: This section would benefit from citations for the statements on impacts of N deposition. Also, this section feels a bit repetitive to the hypothesis section – perhaps those could be combined? In that case, the introduction might benefit from an earlier paragraph wholly focused on the detailed microbial mechanisms that could underly the soil C responses, which could nicely lead into why models like MIMICS and CORPSE aren’t sufficient for evaluating those mechanisms and FG-CORPSE is.
Figure 1: This is a pretty figure, but I think would benefit from some more detail in terms of the connections to CORPSE. It is my understanding that Figure 1 only represents unprotected SOM right? Could that be shown more clearly?
Figure 2: It seems the hypotheses are only about primary decomposers. Would we not expect changes in necromass or DOM under N addition? As mentioned, this should be better justified.
Lines 218-220: I am confused by these lines – I thought the paragraph above was saying there were preferred SOM types based on the microbial functional group? Could the authors clarify the differences between these paragraphs?
Line 226-227: Could the authors clarify in the text – how are these validation data different from the calibration data?
Line 236-237: The same data were cycled over 250 years? Could the authors provide more detail here?
Lines 247-253: Perhaps the three experiments could be numbered here and given a title, so they are easier to track for the reader. Could the authors also elaborate on how the specific parameter changes were chosen? Additionally, I am a bit confused about these experiments. Are they meant to evaluate how the functional groups respond to different litter chemistry to highlight the impact of including CAES in CORPSE? If so, maybe state that really clearly in this paragraph.
Line 257 and beyond: I suggest just using N deposition or fertilization experiment here, instead of the watershed name which is less intuitive to readers. Same suggestion for the control.
Line 292: This is a bit confusing – I think I understand that the authors are modeling recovery, but doing so with baseline parameters or the four hypothesized parameter changes? Could the authors clarify? Calling the experiments the “four N deposition experiments” is largely what is confusing me.
Lines 314-315: I see these numbers as pretty minimal changes – is that how the reader should interpret these or are there notable impacts of these small differences in microbial community?
Line 323: I would say “potentially” demonstrating the need – it could be other missing mechanisms.
Figure 4: I think it might be more logical for the reader to label these downward and then to the next column, since the N deposition data is all covered before the recovery data. Also, adding percent change values, and the label of total “soil” carbon to the C and D graphs would be helpful, as would adding the observational data (for A-D), so the reader can visually compare to observed outcomes.
Lines 399-401: This is true but it is a bit of a shame to be missing the counterfactual of just running CORPSE – It is not really clear if adding the CAES model is what made this work or if it could have been accomplished with just CORSPE alone and it was the way N addition was applied to the model that mattered most. I think it would be wonderful if the authors could run CORPSE alone under these same conditions but given that would be a significant amount of additional work, perhaps the authors could address this possibility that the way N addition was implemented may be more important than the microbial group representation per se. Oh, actually, maybe this is somewhat represented in Figure S1, although seemingly with FG-CORPSE. Still, this would be super helpful to integrate into the manuscript! This would support the idea that microbial functional changes are needed much more strongly than in the current version of the paper, in my opinion.
Section 4.1: As noted above, these community changes are pretty minimal – could the authors comment on why that might be and if these small changes are important for C cycling?
Line 418: Minor, but change to soil C or SOM decomposition, soil in total doesn’t decompose.
Line 459: I would tie reduced CUE more directly to less biomass, necromass, and consequent contribution to SOM.
Section 4.4: This paragraph is fine, but I think could provide more broad recommendations for the community – to me this suggests that microbial adaptation (a la Abs et al., 2025) might need better representation in models.
Abs, E., Saleska, S.R., Allison, S.D., Ciais, P., Song, Y., Weintraub, M.N., Ferriere, R., 2025. Microbiome Adaptation Could Amplify Modeled Projections of Global Soil Carbon Loss With Climate Warming. Global Change Biology 31, e70301.
Citation: https://doi.org/10.5194/egusphere-2026-3566-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 189 | 60 | 21 | 270 | 22 | 21 |
- HTML: 189
- PDF: 60
- XML: 21
- Total: 270
- BibTeX: 22
- EndNote: 21
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Integrating nutrient cycling dynamics and associated shifts in microbial community composition into Soil Organic Matter models is important to capture the complex interactions between soil nutrient and carbon dynamics and to advance our understanding of how changes in land-management affect soil carbon storage in the future. The authors address this challenge by adapting the existing CORPSE (Sulman et al., 2014) model to harbor three instead of just one microbial group (Functional Group-CORPSE) and define three distinct hypotheses of how nitrogen (N) deposition affects one of the newly defined microbial groups. The different hypotheses are then tested in scenario simulations of a 30-year N fertilization experiment and compared to corresponding catchment-scale observational data from the Fernow Experimental Forest. While the paper addresses an important and timely topic, there are several major concerns:
It is not clear how or whether at all nitrogen dynamics are represented in the model. The authors describe the use of the CORPSE model (Sulman et al., 2014) as the basis for the development of their functional group model and to test hypothesis about the effect of N deposition on microbial groups. However, the used version of CORPSE does not represent nitrogen dynamics and there is no indication suggesting that this was added to the model by the authors. This hinders a mechanistic representation of a microbial physiological response to N deposition. Instead, the hypothesized microbial physiological responses are implemented in an ad hoc manner by prescribing defined changes in individual parameter values for times when N deposition occurs. Thus, the model is not mechanistic and applicability to other scenarios is limited (how would microbial functionality respond to half, double, … the prescribed N deposition rate?). Beyond this limitation, it is not clear how robust the presented results (and hence conclusions) are to small variations in the altered parameter values that define the different scenarios or to the assumptions made in the implementation of the different hypothesis (see further comments below). Considering the large uncertainty in microbial parameters, the small differences between the tested scenarios (i.e., small changes in the numeric values of individual microbial parameters) are difficult to interpret mechanistically without additional extensive checks for robustness.
Previous developments of the CORPSE model have already included nitrogen dynamics as well as a coupling between soil microbial and plant dynamics (FUN-CORPSE, Sulman et al., 2017). Both these developments are immediately relevant to address the question raised here. In a recent publication (Ridgeway et al., 2026), FUN-CORPSE was indeed successfully used to generally capture the observed responses of soil carbon stocks to N deposition in seemingly the same (yet more extensive) dataset from the same experiment (30-year N fertilization in the Fernow Experimental Forest). In their study, the baseline implementation of FUN-CORPSE could reproduce the general effect of N deposition on soil C (increase in soil C and decrease in microbial biomass following N deposition) without invoking any of the processes tested here. While Ridgeway et al. (2026) acknowledge limitations in the representation of microbial community processes in their model, their results challenge the robustness of the presented results and conclusions.
Further comments:
It is not clear which and how model parameters were calibrated and how initialization was performed. How robust are the modeled microbial distributions to changes in initial conditions? Do they always converge to the same relative abundances or are other states possible? How would this affect results?
I struggle with the interpretation and implementation of some of the tested hypothesis. For instance, in H1 it is hypothesized that N deposition reduces 1° decomposer abundance, which was implemented by increasing the turnover rate of 1° decomposers when N deposition was active. While this forces the intended effect, it does not correspond to the hypothesized mechanism (reduced exudation of labile C from plants and hence reduced substrate supply for decomposers) but instead could be interpreted as a nitrogen-toxicity effect on 1° decomposers (i.e., an active dying of the organisms rather than reduced growth). In fact, all simulations show the hypothesized decline in degrader abundance making this process redundant. Decreased decomposer function (H3) is motivated by declining enzyme production, yet I wonder if this process can be regarded in isolation or whether this should be accompanied by an increase in CUE: microbes might use the resources they no longer invest into enzyme production to grow more efficiently? Instead, reduced CUE as in H2 might be associated with overflow respiration of carbon under N limitation rather than N addition (Manzoni et al., 2021). Generally, substrate and microbial stoichiometries might play an important role in regulating the microbial response to N deposition which could be considered more explicitly in the tested mechanisms.
A comparison of the here developed Functional Group-CORPSE to the base implementation of CORPSE would be helpful to understand the effect of including the functional groups. As the base implementation of CORPSE (Sulman et al., 2014) aims to implicitly capture a shift in microbial functional groups with a variable microbial turnover rate it would be interesting to see how this compares to an explicit representation of microbial functional groups.
A small mistake in Eq (1): as currently written, sum(M) would cancel in the last right-hand side term.
References
Manzoni, S., Chakrawal, A., Spohn, M., and Lindahl, B. D.: Modeling microbial adaptations to nutrient limitation during litter decomposition, Front. For. Glob. Change, 4, 686945, https://doi.org/10.3389/ffgc.2021.686945, 2021.
Ridgeway, J. R., Sulman, B. N., Weber, S. E., Juice, S. M., and Brzostek, E. R.: Microbially mediated nitrification improves modeled temperate forest responses to declining nitrogen deposition, Appl. Soil Ecol., 217, 106585, https://doi.org/10.1016/j.apsoil.2025.106585, 2026.
Sulman, B. N., Phillips, R. P., Oishi, A. C., Shevliakova, E., and Pacala, S. W.: Microbe-driven turnover offsets mineral-mediated storage of soil carbon under elevated CO2, Nat. Clim. Change, 4, 1099–1102, https://doi.org/10.1038/nclimate2436, 2014.
Sulman, B. N., Brzostek, E. R., Medici, C., Shevliakova, E., Menge, D. N. L., and Phillips, R. P.: Feedbacks between plant N demand and rhizosphere priming depend on type of mycorrhizal association, Ecol. Lett., 20, 1043–1053, https://doi.org/10.1111/ele.12802, 2017.