This is a very well-written article presenting model simulations of Arctic Ocean plankton changes on seasonal to interannual timescales and the implications for potential fisheries. The authors use a recently published, relatively complex, ocean biogeochemical model to assess multiple aspects of Arctic biogeochemistry across timescales. Their findings are well-grounded in past studies and offer a complementary view of the drivers of variability in Arctic Ocean biogeochemistry.

I have a few minor critiques of the article which will hopefully improve it. These mainly relate to additional model validation and more accurate description of the model results. Subject to these changes, I am happy to recommend publication.

Minor comments:

L120-125 It would be useful to also detail how sea-ice influences light limitation in the model. Is NPP permissible under sea ice? If so, does this depend on thickness, snow cover etc?

L161-163. I don’t understand this. Do the authors mean that atmospheric greenhouse gas concentrations are held constant in these simulations so the impact of acidification for example is not simulated?  The prescribed CORE-II fluxes (e.g. heat and freshwater fluxes) are presumably affected by anthropogenic climate change. Or have these forcings been modified in some way?

L202 Is NO3 the only limiting nutrient in the Arctic domain or the only nutrient that is assessed?

L270-272. It is difficult to reconcile this text with what I can interpret from Figure 3. Few if any of the subregions seem to exhibit peaks in observed and simulated chlorophyll in the same month. In the East Siberian Sea, Chukchi Sea and Beaufort Sea the simulated peak appears to be 2-3 months later. While in the Barents Sea the seasonal cycles of observed and simulated chlorophyll almost appear to be anticorrelated. What explains the winter peak in observed chlorophyll in the Barents Sea? Presumably there is insufficient light availability to sustain this? Is this an artefact of variable observational coverage? In which case it might be best to only do pairwise comparisons of models and obs.

Figure 5. I’m surprised that in the Central Arctic light limitation isn’t more extensive in summer. It would be useful to add an evaluation of simulated sea ice extent/thickness. Maybe in Figure 2. Overestimation of seasonal sea ice variability might help explain this and as the authors mention in their discussion, this is an issue that has been previously identified with CORE-II forced simulations.

L365 does “added” just mean simulated here?

Figure 8 is quite difficult to interpret. I suggest avoiding the repetition of labels to allow you to increase the figure size. Why does summer sea ice not appear to be declining? Is this a shortcoming of the simulation setup as mentioned in line 478?

L433-437 It’s not entirely clear to me why these regions are behaving differently. Can the authors expand a little here? Under diminished NO3 one would typically expect fewer large phytoplankton. Why is this not occurring in the Nordic Seas?

This paper presents a relatively new model configuration with many phytoplankton and zooplankton functional types and evaluates seasonal and interannual changes in plankton biomass and community structure from 1948 to 2009. Seasonally, they evaluate changes in phytoplankton biomass and largest limitation terms, quantify zooplankton grazing. When specifically identifying years with low sea ice and high ocean temperature, they find that phytoplankton and zooplankton biomass is higher, plankton phenology was shifted earlier, and potential fishery production was greater. This paper is very well written and provides some valuable modeling insights into possible changes in the future Arctic Ocean, and I have recommended only minor changes prior to publication.

Minor comments:

Chl validation: Here, modeled Chl is compared to satellite-derived Chl using a global algorithm. The Arctic has relatively unique ocean optics, and global ocean color algorithms do not typically replicate Arctic Chl well, so I’d recommend using an Arctic-specific algorithm. This will likely lead to a reduction in satellite-derived Chl, making this comparison look far better, and will also likely shift the seasonality of the phytoplankton blooms earlier.

Nutrients: When nutrient limitation is discussed, does this refer exclusively to NO3 limitation? I imagine that diatoms are limited by Si at least seasonally or in some parts of the Arctic. A little more clarity about what nutrients limit phytoplankton growth would be appreciated.

Results and Discussion overall: I think this section would benefit from greater contextualization with/ comparison to previous studies – perhaps a few sentences in each section. For example, you describe how phytoplankton biomass shifts from largely dominated by diatoms to dominated by smaller functional types as nutrients are drawn down. This is common in global oceans, but has it been observed (or found in other model configurations) in the Arctic? What about the seasonal succession of zooplankton you observe? Similarly, discussion of the predominant limitation terms for phytoplankton growth in other models (or in CESM-1 with Krumhardt et al., 2020, for example) would be valuable. For the fisheries production results, it might be useful to look at other modeled estimates of fisheries changes in the future (e.g. Tai et al., 2019). While this model is only run until 2009, your results about how biomass changes under low sea ice and warmer ocean temperatures suggest a more productive Arctic in the future. Is that consistent with other model findings? A few sentences about these results will better allow readers to assess which of your findings are new contributions and which are consistent with other observational studies or previous modeling studies, giving us confidence in this model configuration.

Typos:

Line 31: there is an unnecessary comma after winter.

Line 114: there should be no parentheses around this citation

Line 146: “62-years” should have no dash

Line 245: 10°C should not have a tilde

Line 277: remove “the” from “the Baffin Bay”

The authors present a numerical study of Arctic planktonic ecosystem productivity and trophic organization in response to major physical drivers, such as sea ice extent, light availability, temperature, several nutrient concentrations, stratification, advection, etc. Their study allows for a detailed exploration of a biogeochemical model of intermediate complexity that considers functional biodiversity and size-based dynamics as well, for a ~60-year time span at the scale of the whole Arctic domain. It leads to a rich harvest of numerical results that are well presented and analysed: despite the ambitious scope of the study, the paper does not feel too complex nor too long. I command the authors for having chosen an interesting balance between a more detailed inter-regional and seasonal approach, and approaches deployed at the larger interannual scale and trophic network level.

With all its qualities, I still think that there are some elements that should be addressed, especially regarding the comparison with observations (validation is too strong a word in the Arctic context).

First, regarding the comparison of simulated and observed surface chlorophyll a level (especially L270-274), I do not agree with the authors that “The comparison between model and satellite chlorophyll shows that in many cases the phenology of chlorophyll, if not absolute magnitudes, corresponded reasonably well”. In most instances, the comparison is not convincing… the only instance where the phenology corresponds are the Nordic Sea (Fig. 3d), while the few instances where the concentrations match are again the Nordic Sea, Baffin Bay (Fig. 3a) and Central Arctic (Fig. 3e) for one month, while Chukchi and Barents Seas are relatively close for June-July, but with and obvious issue for the latter (increasing observed values in winter and fall…). I also stress that in all the chlorophyll a figures a log scale has been chosen, which obviously tends to downplay the discrepancies.

It does not mean at all that the modelling study is not valid, since the model behaves logically and that satellite-based observations of Arctic chlorophyll a are notoriously challenging to handle. I think though that the authors should not downplay this difficult comparison; I’d rather prefer them to recognize it and try to refer to other studies that might have run into the same issue (e.g. Popova et al. 2012. 10.1029/2011JC007112; my reference here is old, the authors might know more recent ones). Moreover, I think the authors could provide an estimation, at least in the discussion, of how the use of a different color algorithm that takes into account the high CDOM content of these waters could change the satellite-based biomass estimates (e.g. Li et al. 2023 10.1364/OE.500340).

Second, starting at section 3.3 approx., it seems to me that about half of the results shown are not / cannot be directly compared to observations, from the slopes of the size-structured community to the trophic position of the zooplankton groups. It would strengthen the paper if the authors could find information on the slope & intercept of the size structure of plankton community, from Tara Ocean, Mozaik expedition or other sources like this, if possible. What is certainly possible, though, is for the authors to present some ideas of experiments that should be done to test the many hypotheses their model has generated! I think it would be very useful for the community of Arctic oceanographers as a whole, since there are so many outcomes of their model that could provide guidance for future in situ experiments.

Detailed comments

1. L76-80: a simple schematic of the different types and groups of plankton would help.
2. L99: “[…] is suited to study Arctic Ocean dynamics.” I think a few more details on issues specific to the Arctic ecosystem, such as usual temperature, light and nitrate limitation should be provided before stating that.
3. L140: while I do not suggest the authors modify the MARBL-SPECTRA model, I would like them to recognize that the choice of a type II functional response is not always optimal and can have a destabilizing influence in NPZD-type models (e.g. Gentleman & Neuheimer 2008, 10.1093/plankt/fbn078; Flynn & Mitra 2016, 10.3389/fmars.2016.00165).
4. L189: why integrating over 150m?
5. L251-253: I am not sure whether he authors speak about a usual spatial trend or a temporal trend resulting from the impacts of climate change?
6. L295: where do the nutrients come from? Remineralization or advection? Both?
7. L451: this first sentence seems to contradict what was just said in the previous paragraph, probably because this model does not take into account sympagic production and its export towards the benthos in spring.
8. L465-466: please provide in one or two sentences indications on how this could have affected your conclusions, much like you did in the following paragraph.