Thank you for your response to your referees. I have reviewed this and have a few suggestions concerning the proposed modifications to the manuscript.

A general remark I would make is that there are several places in your response where you cite either unpublished or grey literature material in support of your changes to the manuscript. I would instead suggest that this manuscript is an ideal venue for such issues to be properly raised, both to support this manuscript and to provide a formally published source for these issues that would be useful for future iterations of this model.

-----

Referee comment: Model has been spun up for three years. Is there any reference to show that this time is enough, particularly for the deeper part of the domain? Or did the authors make some test to check this? (page 8405)

Authors’ response: This was tested in an earlier version of the model and this is described in Hansen (2008), where it was shown that the model did not have any significant drift beyond the first three years. Reference: Hansen, C. (2008): Simulated primary production in the Norwegian Sea – Interannual variability and impact of mesoscale activity, PhD Thesis.

Changes to the manuscript: The reference to Hansen (2008) has been included in the text.

Editor’s response: The referee’s original comment is a good one, but I do not think that the authors’ response is sufficient. It is unclear that the question has been fully answered, particularly in relation to the deep domain. Furthermore, a PhD thesis is not a good source for this sort of information, in part because it is not necessarily easily accessible (I have not checked this) and in part because it is grey literature (i.e. has not received the same degree of scrutiny as a published paper). Far better if the authors just include some evidence in the manuscript to support their contention that 3 years is a sufficient period of time for spin up. For instance, something to clarify what “no significant drift” means and would indicate how much the state of the model is changing with time (e.g. nutrient profiles, etc.). Among other things, this would provide a published source that the authors could use in the future.

Authors’ response: For nutrients this is where we see the improvement and although the bias is better at depth, we do not see much difference between the model performances for different parameters

Editor’s response: Per the previous point, could it simply be that the deeper part of the domain is less perturbed by the biogeochemical model? And on the point concerning a lack of response despite parameter changes could, again, this have something to do with the short duration of the spin-up? Biogeochemical fluxes are typically smaller at depth, so longer spin-ups are can be required to determine whether the model is in “steady state”. Please ensure that the text includes such caveats where appropriate.

Authors’ response: In addition Figures 9 & 10 has been updated to include phosphate and nitrate as the reviewer suggested.

Editor’s response: Figure 9 is impenetrable. It might be preferable to show the model-observation differences as column plots in which the x-axis is month (4 columns) and the y-axis is depth (n rows), with a blue-red colour scale per Figure 5 (i.e. underestimates blue, overestimates red). This would have its own problems (not least that you’d need separate panels for the original and new simulations), but as it is very difficult to make any sense from the plot as it currently appears, an alternative is worth trying.

Authors’ response: Figures R1 and R2 are for the depths 100 to 500 meters, for the depth 500 to 1500 meters there was almost no differences between the runs. Part of the good results at depth is probably due to good initial conditions and relaxation at the boundary.

Editor’s response: Regarding Figure R1, ensure that you comment on whether such a sporadic pattern of skill / no skill could just be random, or whether it reflects real skill in the regions identified as such. Where neighbouring grid cells exhibit very different levels of skill, rather than showing a continuous gradient of skill, one suspects that skill is, in reality, low. Models can be right for the wrong reasons some of the time. Again, material added concerning spin-up

Referee comment: In section 4.2 authors state that changes in mortality of zooplankton produced little effect, contrarily to the expectation but they did not provide any potential reason for this.

Authors’ response: We feel that this is speculative and prefer not to write a possible reason into the paper, but rather say that we are not sure about the reason.

Editor’s response: First of all, were the relative sizes in the parameter changes consistent? It looks to me like grazing was altered to a much greater extent than was mortality (in percent terms). Secondly, the mortality term is described as quadratic but appears to be something different in Table 2. Thirdly, this term has an additional coefficient, k6, whose value likely has a bearing on the insensitivity, but this is not discussed at all. Fourthly, the authors present no information at all on the behaviour of the modelled zooplankton – were they insensitive to the change, or was it just the error statistics?

Authors’ response: Since we do not have any publications showing the (mis)placement of fronts, but we know his is true, we added ‘not shown’ to the manuscript. With regards to a late development of MLD leading to a late spring bloom, we have used Samuelsen et al. (2009) as a reference.

Editor’s response: Perhaps the authors could consider adding such material here? It seems highly pertinent given that the authors are trying to validate the model. Note that there is scope for adding it as supplementary material should it not sit easily in the text. Adding “not shown” is not a useful response to this question.

Authors’ response: ‘not shown’ has been added behind the sentence “The model is consistently late in its initiation time and none of the parameter alterations significantly affected the timing of the spring bloom” to indicate that we have actually looked at this. When the same thing is mentioned earlier in the manuscript, the reference to the paper Samuelsen et al. 2009 has been included. This paper shows the timing issue quite clearly in figure 3.

Editor’s response: See previous comment. Saying “not shown” is not unambiguous nor is it helpful here. Please consider adding the material you refer to as either supplementary material or directly to the manuscript. Incidentally, presumably when you say “figure 3” in your referee response you mean Figure R3? Will this be included in the final manuscript?

Authors’ response: The new formulation is as follows “Not all biogeochemical processes in the ocean are well understood and therefore biogeochemical models are less accurate than circulation models both with respect to model formulations and parameterizations. Observational data for validation and model evaluation are more scarce than for circulation models.”

Editor’s response: Wearing my pedant’s hat I would rephrase this to “Biogeochemical processes in the ocean are less well understood than those of physics, and model formulations and parameterisations are correspondingly less well developed or constrained. Additionally, relative to ocean physics, there are fewer observational data for the validation and evaluation of modelled biogeochemistry.”

Authors’ response: The sentence has been changed to : “Hereby the runs with the original parameter set for both resolutions show no skill for the ME (figure 4) and large negative percentage biases (figure 5), meaning that the model consistently overestimates the chlorophyll.”

Editor’s response: “Hereby” makes no sense; I think you mean “Here”

Manuscript text: “Analysis have shown that the model runs are consistently late in the spring bloom, a persistent feature in this model system (Samuelsen et al., 2009b).”

Editor’s response: Need to pluralise to – “Analyses have shown …”

Authors’ response: The sentence was changed to “The zooplankton mortality is the closure term in the model, but contrary to other studies (e.g (Steele and Henderson, 1992) perturbations of this parameter had little effect on the results.”

Editor’s response: Steele & Henderson (1992) were concerned with how model behaviour changed in response to different functional forms of the closure term, not strictly changes in the associated model parameters (e.g. the maximum loss rate). You seem to be conflating functional forms and parameters here. Please tighten up your wording so that it is clear what you’re referring to and whether your sensitivity test is congruent with Steele & Henderson (1992).

Authors’ response: (7) Phyto-convection is the early seeding of the spring bloom by phytoplankton that was mixed down during winter.

Editor comment: The modification made in the text does not convey this information. I misinterpreted this statement to refer to bioconvection. Please rectify this for clarity.

Authors’ response: (7) We have gone through the manuscript and checked out uses of data, it has been changed in the figure labels as well.

Editor comment: A commonly used convention is that observational science produces “data” while modelling science produces “output” (or “model data”, though this is clunky).

Editor comment: The code availability section usually appears towards the end of a manuscript. Please examine recently published GMD papers for guidance (I should have amended this previously).

Editor: Thank you for your response. There are still a few issues that I think need to be dealt with.

Reply: You are right, we are using a lot of grey litterateur. We have located the ‘grey’ literature that we cite and evaluated in each case if it is possible to use peer-review literature instead. It seems though, that especially for the datasets and models from before journals where available where one could publish such descriptions it is hard to avoid citing reports. We also made sure that this literature is available online.

Editor: While my concern was more to do with the use of grey literature to support statements that attracted the attention of referees, I appreciate the efforts to which the authors have gone here to address this.

On a particular point that might be helpful, I would suggest that the authors consider uploading the Skogen and Søiland user’s guide as supplementary material if possible. This follows the suggestion made in the abstract of GMD’s 2013 editorial:

http://www.geoscientific-model-development.net/gmd_journal_white_paper.pdf

Reply: The thesis is not very specific about the detail of the adjustments over the first three years and mostly describe the drift of nutrients, therefore we have …

Editor: If I am reading this correctly, this analysis has focused on the drift of basin-averaged nutrients. If the basin is large enough, or the input / output fluxes small enough, this sort of drift is always going to be small in size. I believe that the referee was more concerned with the behaviour at smaller scales, for instance for particular water columns. At such scales, there could be large interannual drifts as the model redistributes its currency around its domain even if the total inventory of the same currency shows almost no change. In such a case, it would be difficult to judge the model as being at “steady state”.

That said, in a spatially-resolved model such as this one, driven under observationally-derived forcing for a relatively small number of years, it is not straightforward to establish “steady state”. However, it should still be possible to show how basin-scale properties (or perhaps those in the sub-regions studied in this manuscript) vary with time. I’m thinking here of primary production, plankton biomasses, surface / deep nutrient concentrations. Time-series plots showing how these integrated or averaged properties “level off” over a few years should give an idea of how close to “steady state” the model is. It should also reassure readers that sufficient time has been allowed for the model to reach a semi-repeating annual cycle in these major properties in the sub-regions of interest.

Reply: “Quadratic, rather than linear, mortality in the phytoplankton was one of the changes that had little effect on the error statistics …”

Editor: Your Figure R4 may show the reason for this – the parameterisations of the functional forms used means that the shape of the response curve is not significantly different for the range of zooplankton concentrations studied. I’ve found this in some of my own work (e.g. Yool et al., GMD, 2011), and I would expect it to be a feature of many models. If you include Figure R4 in the manuscript (as supplementary material?), you could modify your text to draw attention to this lack of strong difference in functional response for reasonable values.

Reply: “The zooplankton mortality is the closure term in the model, but contrary to other studies (e.g Steele and Henderson, 1992) perturbations of the functional form of the mortality in N04 had little effect on the results”

Editor: See my previous remark. If you include Figure R4, this response can be strengthened somewhat.

Pg. 15: Change this text “We have shown that the model reproduces a reasonable annual cycle, but one persistent challenge the initiation time of the spring bloom is later than the observations” to “We have shown that the model reproduces a reasonable annual cycle, but the initiation time of the spring bloom is consistently later than the observations”

Pg. 16: Change this text “(the early seeding of the spring bloom by phytoplankton that was mixed down during winter: Backhaus et al., 1999)” to “(the early seeding of the spring bloom by phytoplankton entrained from the shallow seafloor by winter convection; Backhaus et al., 1999)”

Figure 10: The second panel of this figure is fairly impenetrable. It looks to me like the months of January, March and November could be removed without impacting the figure at all. It might also be an idea to retain the colour scheme from the first panel, and show the months of May, July and September as dotted, dashed and solid lines.

Thanks again for your edits and your patience. I have some very trivial amendments to the revisions you have made. I have put comments on your response and attach this here. Please do not hesitate to contact me should you have any further questions.