specific comments



l 18: "The absorption coefficient of CDOM, [...], is measurable in situ and remotely": Can it indeed be measured reliably remotely and distinguished from other (living) light absorbing constituents?



l 40: "the spatio-temporal dynamics [...] reflects the functioning of the carbon cycle": What exactly is meant here, I would suggest rephrasing.



l 41: "non-absorbing": I would suggest changing this to "not light absorbing".



l 50: "There are several possible factors determining...": "Besides the presence of CDOM, there are other possible factors influencing..."



l 63: Since vertical transport is mentioned here, why not horizontal as well? Maybe change  "terrestrial and atmospheric inputs" to something like "terrestrial and atmospheric CDOM sources and physical transport"?



l 73: Is the NW Mediterranean Sea the study region?



l 75: "depositions from the atmosphere are 2–5 times larger in the Mediterranean Sea than in the oceans": higher than "in the oceans" is not very specific: higher than anywhere in the ocean, higher than some average?



l 85: What does "probably controlled" mean? Perhaps change to "mostly controlled" or "its dynamics are dominated by"?



l 86: One can maybe make the counterpoint that photoacclimation and nutrient limitation are also often controlled by vertical transport. Maybe distinguish between being controlled by and directly caused by, as the preceding sentence suggests?



l 92: "observations in the NW Mediterranean Sea suggest direct CDOM production by phytoplankton": Is this then equivalent to the way it is handled in other models, i.e. the fixed fraction of the dpp of phytoplankton?



l 169: Explain the symbols and superscripts used here ("O^{(3)}" etc.).



l 171: "A subscript appended to each module ...": This sentence requires more explanation. What are the modules? Is Chl-a considered an elemental constituent? I think I understand the rest of this paragraph, but this sentence is difficult to follow and seemingly unrelated to the preceding and following sentences.



l 184: What is "OAC"?



Eq 1: "; i = 1 to 4" -> " for i = 1,2,3,4"



Eq 1: Why does zooplankton indexing start at 3?



l 191: As an aside, I find the "prd" abbreviation of "predation" slightly confusing, as I always think of "production" first. I would suggest using "grz" for "grazing" instead.



Eq 2: Underneath, it reads like both temperature and nutrients are dimensionless, when really these temperature- and nutrient-dependent factors are dimensionless. I would suggest changing "ft_P" to "ft_P(T)" where T is model temperature. The same applies to fn_P, which has the benefit of telling the reader which nutrients are entering this equation and control GPP. Currently, it might appear to a reader as if GPP is independent of other model variables.



l 195, 197: Sometimes "mg Chl", sometimes "mg Chla" is used. To stay consistent with previous notation, I would suggest using "mg Chl-a" and "(mg Chl-a)^{-1}" when in the denominator.



Eq 4: I presume, P_P is the phosphorous quota (subscript P) of phytoplankton (regular P). Is the superscript denoting the phytoplankton type missing, or am I misunderstanding? Similarly, is the "P_N^{(k)}" meant to be "P_{N^{(k)}}"? That is, the superscript "(k)" appears to be attached to the P, but should it be attached to the subscript N instead?



Eq 5: What is i set to here, is there perhaps a sum missing?



l 231: Why do the other zooplankton not excrete labile DOC? Are they too small/large? This information is missing from the manuscript.



Eq 14-16: Out of curiosity, why are different reference wavelength used in the 3 equations (450 nm, 440 nm and 550 nm)?



l 356: "()" are missing for a_PH and a_DG.



Fig. 2: x- and y-axis labels are missing. Consider coloring land gray in the top left map.



l 378: "assumed equal": Why "assumed"?



l 389: "and restored at monthly frequency": Does this imply nudging/relaxation?



l 419: I would suggest removing the "while" here.



l 425: What does "assumed true" mean, assumed that the true value is known and not included in the optimization? I would prefer "assumed known" or "assumed constant". Please make it explicit that this implies that the parameter in question is not included in the optimization. This needs to be made clear initially, the phrase is used a lot.



l 426: Subscript "(5 to 6)" and "(1 to 4)" are unconventional and could be confusing some readers. I suggest writing them out fully, i.e. (I am not including superscripts here): f_Z(5), f_Z(6), f_P(1), ..., f_P(4), and f_B.



l 432: "All optimized parameters appear in some process independently:": As a reader, I was now expecting a brief review of where the parameters mentioned in the previous sentences appear in equations. Instead, new parameters are mentioned and added to the list, of which two are not actually appearing in some process independently. Does the "all parameters" include the previously named ones? It would be helpful to break up this long sentence into two, don't use "all parameters" (one could use "the following" or similar).



l 435: "We assumed true a_PS and optimized \phi_C, because these parameters are not well documented in literature being more difficult to measure.": Does this mean \phi_C was included in the optimization and a_PS was not, because they both appear in the same equation, and they may be dependent on each other?



l 441: "A total of 25 optical and BGC parameters were optimized": I think it would be helpful to mention this at the beginning of the paragraph.



l 449: "O_1, ..., O_n, O_N": This notation is confusing, why include n and N here, is n=N-1? The same applies to P.



Eq 27: Similar confusion, I assume n iterates from 1 to N here. So it should be "n = 1, ..., N" or (more conventional) "n=1" and the "N" is placed on top of the sum. The same applies to equation 29.



Eq 29: So N_m is the number of observations for variable m? This needs to be made explicit, and I suggest introducing this notation right away in Eq 27, so that it is clear what SSQ_m is. I probably do not understand the terms here and what the subscript m indicates, it appears like SSQ_m/(2*\sigma_m) = SSQ_m/(2*SSQ_m/(N_m-1)) = (N_m-1)/2? This is very confusing.



l 460: The DE algorithm is not explained very well. For example, it is not clear from the text which of the symbols denotes the target vector and which is the mutant vector. What does "crossing" with the mutant vector mean? In "All target vectors in the first generation are crossed with the mutant vector", it sounds like the same mutant vector is applied to all target vectors. If I did not know how DE worked, I would not have a better idea after reading this section. This description needs to be rewritten.



l 466: "optimal set of 25-parameter values": The DE algorithm aims to find optimal parameter values, but that does not mean that the resulting parameters are indeed optimal in that they maximize the likelihood. Hopefully, they are close to the global optimum. In other words, I would suggest using "estimate of the optimal parameter values".



l 476: "reversed some of the assumptions of the optimized configuration": What does "reversed" mean here? I think "changed" might be more suitable.



l 479: "The comparison of the results of Optimized with Nutrients and Constant constituted and experiment..." This sentence is difficult to follow, I suggest rephrasing. I had to read the previous sentences several times as well: one uses "removed the dependence >>of<< f on \theta/\theta_chl", the other "removed both dependencies >>from<< f, on \theta/\theta_chl", making it very difficult to understand what is dependent on what, and what is being removed. Here, I strongly suggest stating the aim of each experiment first and then go into details of the implementation.



Fig 3: Mention that these are climatological values in the caption.



Fig 3: Units in the color bar and in the panel are not identical (they are equivalent, but why create confusion).



l 568: Out of curiosity, what are the correlations between in situ and satellite observations?



Fig 7: What happened in the model in early 2013 when there appears to be a drop of model chlorophyll to nearly zero?



Fig 8 and 9: it might be helpful to change the order of legend entries to match the extra axes.



Fig 11b: Why are the ranges of reported values so patchy?



l 683: "1) depends": Missing "it".



l 684: "Two additional simulations considered an average value of 6.2% of CDOM in the leakage flux": writing "considered an average value" is not very precise, the reader needs to know exactly what is happening, for example: "Two additional simulations use a constant fraction of CDOM in the leakage flux, setting the CDOM fraction to 6.2%, which is the average value obtained from the Optimized simulation."



Fig. 12: Panel (d) says "0-9m" but (e) is missing units: "40-60". It would also be useful to have this information in panels (b) and (c).



l 799: "This is in better agreement with observations when the proportion of coloured material in the dpp depends on nutrient and light constraints": Add a reference to the model here, and change the "when" to "if", otherwise readers might think the "when" means "at times when" which would completely change the meaning of the sentence. Proposed new version: "This is in better agreement with observations if the proportion of coloured material in the dpp is modelled to depend on nutrient and light constraints"



l 809: Light limitation of what?



l 836: "Adding optical components to BGC models greatly improves the comparability of simulated and observed types of variables": What exactly is meant here, it becomes easier to compare model and observations when the model has optical components?