| 24 Jul 2026
ECOSMO E2E v2.0: Generic modules of higher-trophic-level fish and macrobenthos functional groups compatible for two-way coupling to lower-trophic-level model hosts through FABM coupler
Abstract. The ECOSMO E2E v2.0 model includes functional-group-type fish and macrobenthos modules that can be two-way coupled to N(nutrient) P(phytoplankton) Z(zooplankton) D(detritus) type lower-trophic level (LTL) models. The model has been reprogrammed from the previous version, ECOSMO E2E v1.0, into independent fish and macrobenthos modules. This new version utilises the conceptual advantage of the Framework for Aquatic Biogeochemical Models (FABM), which provides an interface for coupling ecosystem models of diverse types. The coupling is exemplified by employing three different LTL models widely used in Copernicus Marine Services applications (ECOSMO, ERSEM, and ERGOM). The flexible design of the ECOSMO E2E v2.0 allows it to be coupled to LTL models and applied to regions beyond the example given in this paper. We present a proof of concept showing the two-way coupling of the E2E v2.0 with these LTL models, implemented in the central North Sea. We use a 1D water column model based on the General Ocean Turbulence Model for simulating the physics. The control simulation includes one macrobenthos functional group and two fish functional groups: a predominantly planktivorous group and a predominantly bentho-piscivorous group. The modelled annual cycle of fish production is consistent with the patterns observed in fish biomass estimates from the International Bottom Trawl Survey in the North Sea. We present a model intercomparison of simulations from coupling the E2E model with different LTL models to explore the ecosystem dynamics of the central North Sea. We assess the effects of two-way coupling between lower and higher trophic levels on plankton dynamics, nutrient cycling, and fish production. Two-way coupling induces stronger and more realistic top-down control on zooplankton and phytoplankton than simplified LTL closures, modifying both biomass and seasonal patterns. Macrobenthos enhance detritus remineralisation, leading to elevated bottom-layer nutrient concentrations. Fish production strongly depends on macrobenthos, and the impacts on LTL fields vary between models according to trophic structure. We show that differences in prey composition among different models further drive non-additive ecosystem responses that shape fish and macrobenthos production.
The manuscript “ECOSMO E2E v2.0: Generic modules of higher-trophic-level fish and macrobenthos functional groups compatible for two-way coupling to lower-trophic-level model hosts through FABM coupler”, by Vijayakumaran et al., describes a rewriting of a first version, ECOSMO E2E v1.0, that permits a modular definition of HTL food webs and that can be coupled with independent LTL models. To illustrate the value of such an update, the authors complete an intercomparison of the model when forced by three different LTL models, ECOSMO, ERSEM and ERGOM, on a virtual 1D profile representative of the central North Sea. Through the comparison and a set of numerical experiments, the authors discuss several aspects of the simulations, including (1) the role of macrobenthos, (2) the value of two-way coupling and top-down control of HTLs, and (3) the non-additive effect of macrobenthos and fish.
The new developments, essentially technical, are here presented as motivated by the need to incorporate two-way coupling between HTL and LTL models. The exercise shows significantly different results between models with or without coupling, and with different modules turned on and off, which the authors discuss in detail. However, the point of the analysis is not clear. First, because comparison to observations is largely qualitative and limited. Modifying the structure of the LTL model and turning on and off modules leads to different dynamics, but beyond being a conceptual advance, how does this demonstrate that two-way coupling improves the model's performance? Second, the analysis relies on a single parameter set — with no clear explanation of how the parameters were selected — so it is hard to gauge the generality of the aspects discussed by the authors. I understand that when comparing the role of coupling to LTL models, a common set of parameters for E2E is the appropriate approach but introducing some variability in the parameter combinations picked is necessary for a more informed analysis of the implications of the two-way coupling, or of the suite of numerical experiments. Otherwise, some of the conclusions are induced by the parameter selection — for example, the essential role of macrobenthos on fish production, which is somewhat unsurprising given that the parameters selected indicate a strong preference of fish for macrobenthos.
For the analysis to be more robust: (1) clarify the purpose of the numerical experiments — if not to disentangle how two-way coupling improves the match with observations — by highlighting robust similarities between LTL models per experiment, and dissimilarities. This seems to be what the authors are attempting, but it could be further clarified, for example with dedicated sections in the discussion, or with paragraphs better justifying why the focus on the role of macrobenthos, or why the test on additivity; (2) the authors should provide more details and references on the parameterization, and ideally explore the sensitivity of the analysis to these parameters; (3) finally, some conceptual choices need to be better explained, e.g., why mortality is not routed to the silicate pool (see details below).
See following more detailed comments
Title: “ECOSMO E2E v2.0” in the title is confusing, as the authors present a model coupled with three different LTL models (ECOSMO, ERSEM and ERGOM). Could you not simply refer to E2E v2.0, as you do in the abstract L9?
L11: “The modelled annual cycle of fish production…”, first production should be replaced by biomass, and fish biomass estimates by CPUE estimates. Then I suggest not highlighting consistency with observation here, given the large variations between LTL models. This reads as overstated.
Eq4, 11: The model includes a temperature dependence of excretion, but not of consumption. Please explain or comment on this assumption, as for ectotherm one would expect temperature dependence of consumption (see Sherman et al. 2016, Englund et al. 2011).
Eq16 : typo? RMBCons -> RMB(cons) ; check also Cons vs. cons across equations.
Eq18 : why is there no conversion factor for nutrients, like there is for O2?
Eq19 : why /dz? Isn’t CMB already in mmolC/m3? Please also mention the units of all variables, this allows better tracking of the equations, and while I understand some are already detailed in tables 1-3, some are not mentioned (like the unit of CMB).
Section 2.1: The parameters of the model are not explained nor referenced; more explanation of these is necessary. Testing the sensitivity to the parameters that are least well constrained would improve the generalizability of the conclusions of the paper.
Figure 1: Why the category user-defined prey listed as an input? Aren’t all categories user-defined in the first place?
L232: “mortality fluxes target NH4 , PO4…” my understanding is that the authors are not directing mortality to silicate, so that the models are comparable to ERGOM, which has no silicate. However, in doing so, the coupling might break the stoichiometry of nutrients in the LTL models, with indirect effects on the primary producers. The authors must add this silicate flux where it applies, or clarify why it is not a problem.
L259: mg m-3 or mmolC m-3? mg seems inconsistent with the half-saturation constants r; please clarify.
Section 3.3: Please list the experiments in order of appearance in the rest of the section, eg “Control Run” should be listed as the first experiment.
Section 3.4: Please add more details here, eg what data? CPUE I assume. Why the conversion to biomass, and how? I understand this is detailed in Daewel et al. (2019), but a couple of sentences here would help readability. Also, at L300, please provide the link to the data, rather than the ICES homepage.
Figure 2: Perhaps plot shortwave radiation on a log scale, as it decays exponentially. This will make the variability easier to see.
L345: “These characteristics are consistent with the model simulations”, please plot the vertically integrated zooplankton seasonal cycle. Based on the authors’ statement, I would expect a doubling from winter to summer; however, looking at the profiles figure 4, zooplankton seem nearly absent in winter, this must be improved or at least discussed.
L353: If I understand properly, only medium-size detritus is shown for ERSEM, and therefore looks markedly lower than the other models. If small-size detritus is much more abundant, and presumably closer to ERGOM and ECOSMO, why not pick the small detritus group for the coupling?
L370: the model shows fish biomass, not “fish production potential”, production would require a flux, ie. per unit of time. Moreover, I understand that observations and simulations are not directly comparable, and I accept the qualitative comparison here. But the lack of a sensitivity analysis of E2E to its parameters limits the scope. It is a 1D configuration that should be reasonably fast to run, why not attempt replicate simulations and communicate results based on this ensemble?
L426: “In the ERGOM-E2E…”, sorry but I don’t understand the sentence. Are you referring to the near-identical ERGOM-E2E curves between the two comparisons? Please clarify.
L428: typo: presenence -> presence
Section 5: Here the authors conclude on the insights of the different experiments, while the introduction was geared towards the value of two-way coupling. It is unclear why the authors focused on these particular insights and not others—for example, low trophic-level diversity, or the presence vs. absence of trophic cascades. The manuscript would benefit of a clearer framing of why these numerical experiments were run, and why the focus on the role of macrobenthos and additivity (or lack thereof) of the HTL groups. Otherwise, the purpose of the analysis is not clear.
References:
Englund, G., Öhlund, G., Hein, C.L. and Diehl, S., 2011. Temperature dependence of the functional response. Ecology letters, 14(9), pp.914-921.
Sherman, E., Moore, J.K., Primeau, F. and Tanouye, D., 2016. Temperature influence on phytoplankton community growth rates. Global Biogeochemical Cycles, 30(4), pp.550-559.