the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 20 Oct 2025
A mechanistic model of hypoxia-driven benthic carbon cycling integrating microbial energetics and faunal mortality
Abstract. Hypoxia reduces the mineralization of organic detritus and increases mortality in benthic fauna, both of which alter carbon storage through complex changes in organic matter and calcium carbonate (CaCO₃) dynamics. To mechanistically assess these processes, we developed a new model that links oxic, suboxic, and anoxic mineralization pathways to microbial ATP production efficiency. This formulation was incorporated into the benthic–pelagic coupled model EMAGIN-B.C., resulting in an extended version designated EMAGIN-B.C.-MR (MR: mineralization rate). The model also includes revised mortality and metabolic suppression functions for benthic fauna under oxygen-deficient conditions and explicitly couples suspension-feeding benthos biomass with CaCO₃ production and burial fluxes. We applied EMAGIN-B.C.-MR to Tokyo Bay, a eutrophic coastal system prone to seasonal hypoxia, to simulate long-term changes in carbon cycling under hypoxic (0 mg L⁻¹) and non-hypoxic (5 mg L⁻¹) summer conditions. Results showed that hypoxia enhanced detritus storage and burial by both suppressing microbial degradation and reducing bioturbation and grazing due to suspension-feeding benthos mortality. Conversely, CaCO₃ production and burial declined owing to inhibited shell formation. These dynamics revealed that total carbon storage is shaped by interacting biogeochemical and ecological feedbacks, resulting in nonlinear trajectories under repeated hypoxic stress over decadal timescales. By integrating microbial energetics and oxygen-sensitive faunal responses, the EMAGIN-B.C.-MR model provides a mechanistic framework for assessing benthic carbon cycling under deoxygenation. This framework offers biogeochemical insights into the regulation of organic and inorganic carbon burial balance by oxygen availability – with implications for coastal carbon budgets, blue carbon management, and climate feedbacks – and is applicable to other oxygen-deficient environments such as eutrophic estuaries and semi-enclosed seas.
- Preprint
(1853 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-4822', Anonymous Referee #1, 10 Nov 2025
-
AC1: 'Reply on RC1', Akio Sohma, 02 Mar 2026
We sincerely thank the referee for the careful and detailed evaluation of our manuscript. We recognize that the original submission did not describe the model structure, parameterization, notation, and dimensional consistency with sufficient clarity, which may have given the impression of insufficient transparency. We have substantially revised Section 2 to provide explicit parameter definitions, dimensional consistency checks, complete functional descriptions, and consistent notation throughout the manuscript. Below we clarify the conceptual novelty of the formulation and address each specific concern raised.
A. Conceptual and structural novelty of the mineralization formulation
In classical redox-dependent sediment diagenesis models (e.g., Soetaert et al., 1996; Sohma et al., 2008), oxygen and nitrate concentrations determine the partitioning among oxic, suboxic, and anoxic mineralization pathways through oxygen–nitrate concentration-dependent switching functions. However, in these formulations, the mineralization rate per unit detritus is prescribed empirically and remains invariant under changes in thermodynamic energy yield. Thus, redox conditions redistribute flux among pathways but do not alter the intrinsic mineralization efficiency.
In contrast, the present study introduces a thermodynamically constrained formulation in which ATP production efficiency (αATP) is derived from pathway-specific Gibbs free energy yields. For example, oxic respiration yields approximately 402 kJ per unit organic matter oxidized, whereas suboxic and anoxic pathways yield 359 kJ and 223 kJ, respectively (Table 1). These differences directly determine ATP generation efficiency and thereby constrain microbial growth potential.
Importantly, the mineralization rate constant is no longer externally prescribed as a fixed empirical coefficient. Instead, it becomes an emergent property derived from energy balance constraints:
- Pathway-specific Gibbs free energy yield
- ATP production efficiency
- Microbial biomass yield per ATP
Accordingly, while the mineralization equation remains mathematically first-order in detritus concentration, its rate constant is mechanistically redefined. Redox-dependent variation in mineralization efficiency arises from thermodynamic energy limitation rather than empirical tuning.
This distinction is important: the model does not merely redistribute flux among respiratory pathways but links mineralization efficiency explicitly to thermodynamic energy availability under hypoxia. Because total mineralization flux is expressed as k × DET, reductions in energy yield under hypoxic conditions decrease k, promoting detritus accumulation and introducing nonlinear feedback between oxygen availability, mineralization efficiency, detritus storage, and burial fluxes. Such feedback cannot emerge when the rate constant is fixed and independent of redox energy yield.
We acknowledge that this conceptual distinction was not sufficiently emphasized in the original manuscript and have revised the text to clarify it explicitly.
B. Clarification of model structure and physical processes
The benthic module embedded within EMAGIN-B.C. represents an early diagenetic biogeochemical framework coupled to pelagic–benthic exchange processes. It includes:
- Organic matter deposition and mineralization
- Redox-dependent respiratory pathways
- Benthic fauna dynamics
- Carbonate production and dissolution processes
- Coupled pelagic–benthic carbon exchange
In the revised manuscript, we have expanded Section 2 to explicitly describe these components and clarify the physical and biogeochemical processes represented.
C. Parameter units, symbols, and dimensional consistency
We agree that insufficient clarity regarding units and symbols hindered interpretability in the original submission.
In the revised manuscript, we have:
- Provided explicit units for all parameters and state variables in Tables 1 and 2
- Ensured consistent notation throughout (e.g., avoiding dual use of T for time and temperature)
- Distinguished clearly between parameters and state variables
- Verified and explicitly stated dimensional consistency
Specifically, the mineralization flux is now explicitly shown to have units of C-mol T⁻¹, and dimensional consistency is stated immediately following the governing equation.
D. Microbial biomass dynamics and mortality
The referee correctly notes that if microbial biomass were treated as an independently prognostic state variable, explicit mortality terms would be required.
We clarify that microbial biomass (MBAC) in the present formulation functions as an intermediate variable linking ATP production efficiency to mineralization rates. It is not treated as an independently prognostic ecological compartment. By combining assumptions (1)–(6), intermediate variables such as microbial biomass and ATP production are algebraically resolved under quasi-steady growth conditions. Therefore, explicit mortality terms are not required within this formulation.
We have revised the manuscript to clearly state this structural role and avoid the impression that full microbial population dynamics are being simulated.
E. First-order structure of mineralization
We agree that the mathematical form of mineralization remains first-order with respect to detritus concentration. However, the conceptual advancement lies in the redefinition of the rate constant.
In conventional models, the rate constant is empirically prescribed and independent of redox-dependent energy yield. In the present formulation, the rate constant becomes mechanistically constrained by pathway-specific Gibbs free energy yield and ATP production efficiency. Thus, although the equation retains a first-order structure, its mechanistic basis differs from purely empirical first-order representations.
We have revised the manuscript to clarify this distinction more explicitly.
F. Oxygen–nitrate concentration-dependent function
The oxygen–nitrate concentration-dependent switching function follows established formulations in sediment diagenesis models (Soetaert et al., 1996; Sohma et al., 2018). In the revised manuscript, we now explicitly describe its functional structure in the main text rather than relying solely on citation, ensuring full transparency of the redox partitioning mechanism.
G. Formalism and notation consistency
We appreciate the referee’s comments regarding symbol usage and notation inconsistencies (e.g., δ symbols, dual use of T). These issues have been corrected to ensure mathematical clarity and consistency throughout the revised manuscript.
Concluding remarks
We sincerely thank the referee for highlighting areas where the original manuscript lacked clarity. The revised manuscript now provides:
- Explicit model structure and process descriptions
- Complete parameter definitions and units
- Verified dimensional consistency
- Clear mechanistic positioning of the thermodynamic constraint
- Transparent formulation of redox partitioning functions
We believe these revisions have substantially improved the clarity and transparency of the manuscript and have addressed the referee’s concerns.
Citation: https://doi.org/10.5194/egusphere-2025-4822-AC1
-
AC1: 'Reply on RC1', Akio Sohma, 02 Mar 2026
-
RC2: 'Comment on egusphere-2025-4822', Anonymous Referee #2, 24 Feb 2026
To thoroughly discuss the paper's topic, "The effects of hypoxia on
carbon dynamics within sediment layers," the following three points
should be considered:
1. Quantitative assessment of DIC caused by detritus mineralization
(organic matter decomposition).
A large amount of CO2 is generated during the detritus decomposition
process. This CO2 converts to carbonate, which reacts with Ca++ ions
to form CaCO3 and precipitate (calcification). To discuss changes in
the CaCO3 process due to hypoxia, we must consider not only changes in
calcium shell formation by shellfish, but also changes in this flux.
The amount of CO2 generated varies depending on the respiratory
substrate used by bacteria and can be quantified for each of the oxic,
suboxic (nitrate reduction), and anoxic (manganese, iron reduction,
and sulfate reduction) processes.
This can be used to estimate the calcification flux, but to do so, it
is necessary to model the five-stage decomposition process that
transitions from oxic to anoxic.
2. Tolerance of benthic communities to reducing substances.
This model only addresses the oxygen tolerance of benthic communities.
The scenario seems to be to examine the impact of hypoxia on the
amount of benthic organisms, which in turn affects the formation of
CaCO3. However, the impact of reducing substances on benthic habitats
cannot be ignored. In particular, the toxic effects of H2S, generated
during the decomposition process by sulfate-reducing bacteria, are
important (as seen in blue tides). Even organisms highly tolerant of
hypoxia may die from H2S, so ignoring this makes no sense. Quantifying
H2S production requires modeling the sulfate reduction process
mentioned in ① above.
③ Modeling the dynamics of oxidizing and reducing substances (modeling
early sediment diagenesis).
The paper compares calculated DO, PP, and PO4-P values in the water
column with observations and concludes that the model results are
valid.
However, PO4-P is underestimated in the summer when hypoxia
progresses, and this is not sufficiently justified. Phosphorus binds
to iron hydroxide and precipitates in sediments. When H2S is produced
in an anaerobic environment, the reaction to form iron sulfide
progresses, promoting phosphorus elution. The discrepancy may be due
to insufficient evaluation of phosphorus elution into the water column
under hypoxic conditions.
In any case, I think the challenge is to more precisely model the
dynamics of materials in sediments, including not only C, N, and P but
also manganese, iron, and sulfur. It would make sense to discuss
changes in carbon cycle dynamics due to hypoxia based on this model.
There is no description of a material circulation model within the
sediment layer. Furthermore, the model results have not been verified,
so this should be described as a future issue.
There are several errors in the manuscript, which I would appreciate
corrections for:
- Line 95: Units of detritus biomass
: Fig. 2: There are several errors in the symbols (e.g., δ is written as C).
Fig. 4: There is an error in the legend (b. and c. are reversed).Citation: https://doi.org/10.5194/egusphere-2025-4822-RC2 -
AC2: 'Reply on RC2', Akio Sohma, 02 Mar 2026
We sincerely thank the referee for the thoughtful and technically detailed comments. We appreciate the opportunity to clarify the structural scope and scientific positioning of the present model.
- DIC generation during mineralization and CaCO₃ processes
We agree that detritus mineralization generates CO₂, which enters the carbonate system and may subsequently influence CaCO₃ precipitation. We would like to clarify that DIC production from organic matter mineralization is fully and stoichiometrically accounted for in the EMAGIN framework (Sohma et al., 2018). Each redox pathway (oxic, suboxic, and anoxic) generates dissolved inorganic carbon (DIC) according to mass-balanced reactions, and these fluxes are dynamically coupled to the carbonate equilibrium module embedded in EMAGIN-B.C.
Therefore, CO₂ production associated with mineralization is explicitly represented in the model, and its effects on carbonate chemistry are internally calculated. The present study evaluates how hypoxia modifies mineralization efficiency and benthic calcifier biomass, thereby influencing CaCO₃ production and burial.
We acknowledge that a fully resolved five-stage early diagenesis model (oxic respiration → nitrate reduction → manganese reduction → iron reduction → sulfate reduction) would provide a more detailed representation of redox transitions. However, such parameter-rich modeling requires extensive sediment geochemical data (Mn, Fe, S species, and porewater fluxes), which are currently not available for Tokyo Bay at sufficient spatial and temporal resolution to constrain a fully mechanistic multi-element reaction network.
The primary objective of this study is to assess hypoxia-induced changes in mineralization efficiency and benthic mortality within a coupled pelagic–benthic–carbonate framework. This scope has been clarified explicitly in the revised manuscript.
- Reduced substances and H₂S toxicity
We appreciate the referee’s comment regarding the potential toxicity of reduced species such as H₂S.
Reduced substances are not neglected in the present model. The sediment module adopts the oxygen demand unit (ODU) concept following Soetaert et al. (1996). In this framework, reduced chemical species generated under anoxic conditions—including contributions from sulfate reduction (H₂S), iron reduction (Fe²⁺), and manganese reduction (Mn²⁺)—are represented collectively as reducing equivalents.
Thus, the accumulation of reducing conditions associated with hypoxia is internally represented. However, the present study does not explicitly model species-specific toxicity thresholds (e.g., lethal H₂S concentrations for benthic fauna). Instead, dissolved oxygen (DO) is used as an integrated ecological stress indicator.
We agree that explicit sulfur and metal cycling with toxicity parameterization would provide a more mechanistic description of benthic mortality under severe reducing conditions (e.g., blue tide events). Such development would represent a substantial extension of the current framework and is identified as an important future research direction.
- Modeling sediment carbon cycling and phosphate dynamics
We respectfully clarify that sediment carbon cycling is explicitly represented in the present model. Organic matter production, redox-partitioned mineralization (oxic, suboxic, anoxic), benthic faunal metabolism, sedimentation, and burial fluxes are dynamically calculated within the sediment module. As illustrated in Figure 9, the production and consumption processes of benthic detritus are resolved at the process level under both hypoxic and normoxic conditions.
The focus of this study is not to present a comprehensive reaction–transport early diagenesis model, but rather to evaluate how hypoxia modifies specific biogeochemical processes and thereby alters carbon storage dynamics.
Regarding phosphate, we agree that under strongly reducing conditions, phosphorus release can be enhanced through Fe–P–S coupling (reductive dissolution of Fe(III) oxides and formation of FeS). The current sediment module does not explicitly resolve the full Fe–S–P reaction network, which may contribute to underestimation of summer PO₄ under severe hypoxia. This limitation has been clarified in the revised manuscript.
It should be noted that the model is fully process-based and does not rely on data assimilation. Despite strong nonlinear coupling among phytoplankton production, zooplankton grazing, oxygen dynamics, ammonium regeneration, and nitrate reduction, the model reproduces seasonal patterns of multiple variables with statistically significant correlations (Table 3), including:
- Bottom DO: R = 0.83
- Surface NO₃: R = 0.73
- Surface POC: R = 0.80
- Surface PO₄: R = 0.66
These results indicate that the dominant coupled biogeochemical processes are captured at the system scale. Further refinement of Fe–S–P coupling would likely improve PO₄ representation under extreme hypoxia and is identified as a future extension of the model.
- Minor corrections
We thank the referee for identifying several errors in the manuscript.
- Line 95: The unit of detritus biomass has been corrected to ensure consistency with the carbon-based formulation (C-mol basis) used throughout the manuscript.
- Figure 2: Symbol inconsistencies (including incorrect notation such as δ and C) have been corrected, and notation has been standardized.
- Figure 4: The legend has been corrected to properly distinguish panels (b) and (c).
All of these issues have been revised in the updated manuscript.
We thank the referee again for the constructive and technically insightful comments. We believe that the clarifications and revisions described above address the referee’s concerns and clarify the structural scope and scientific positioning of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2025-4822-AC2
-
AC2: 'Reply on RC2', Akio Sohma, 02 Mar 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 198 | 52 | 21 | 271 | 14 | 18 |
- HTML: 198
- PDF: 52
- XML: 21
- Total: 271
- BibTeX: 14
- EndNote: 18
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
In this manuscript, the authors apply an *improved* model of benthic carbon cycling to estimate sedimentary carbon cycling and storage in Tokyo Bay. The authors claim to present a novel formulation of detritus mineralization that is based on ATP production efficiency of microbes performing oxic, suboxic and anoxic mineralization. This they argue is an improvement over other formulations that assume simple first-order kinetics.
While the paper makes a lot of claims to improve a model, it is very difficult to assess whether this is so. First of all, the model itself is very crudely explained, if at all: one has to assemble bits and pieces to deduce that the benthic model is probably a biogeochemical (early diagenetic?) model, but it is not stated which processes this model describes (e.g. which physical processes?). Units of parameters are rarely given, the parameter values are lacking, tables do not use proper symbols for parameters or variables, and so on. With so much information lacking in this manuscript, it is difficult to interpret the results.
The text has many inconsistencies. For instance, it is said that the model introduces ATP production rates and microbial biomass dynamics, so one would assume that microbes must die else their biomass would only increase - but this is unexplained. Also, based on the Figure 2, it appears that bacterial biomass is set as a fixed fraction of total detritus instead (equation 5), but how this is deduced is unclear.
From figure 2, equation 5, it appears that in this *new* formulation the detritus mineralisation is (again) first – order to detritus. So, as far as I can see, the new formulation is the same as the formulation that it replaces, i.e. first-order mineralisation. What is then the point?
The formalism is very non-standard, e.g. the use of the delta in figure 2 (to represent a derivative or what?). There is inconsistent use of symbols (e.g. T which is used for time is also used for temperature. Many things are unexplained, e.g. what is the “oxygen-nitrate concentration dependent function” in table 1, and so on. Based on all this, the underlying manuscript is not of sufficient quality to be admissible for publication.