Dear authors,
The topic of your manuscript has been evaluated as of significant scientific interest and the overall presentation of good quality. However, both reviewers have raised some issues that I would urge you to constructively attend to, in particular, regarding the interpretation of DCM patterns produced by model simulations and how other processes (i.e. nutrient dynamics, grazing...) not discussed in the text affect patterns observed.
Sincerely,
Christine Klaas

Dear authors,
Based on commentaries by the reviewers, your manuscript can be accepted for publication pending some minor changes. In particular I refer you to the points 1-5 raised by one of the reviewers (see below). I am also attaching a annotated manuscript with suggestions for improving the manuscript.
Sincerely,
Christine Klaas

1. One of the declared scopes of the paper (e.g., Title, p.1 l.5, p1 l.23-p.2 l.7) was to explore the advantage of assimilating Argo profiles in a coupled model. Reading the manuscript I understood that the assimilated variables were, in turn, PAR and chlorophyll a profiles to compute PAR with several bio-optical models. As for the mixed layer depth, which is one term to modulate the diffusivity profile, it is not clear if it responds to external forcing, e.g., p.5 l.5, or is also an assimilated variable. I might have missed this. If not, this should be clarified.
2. On the other hand a significant part of the text is devoted to discuss not too much the feasibility of ARGO data assimilation in a coupled model but the mechanisms determining the DCM dynamics. This is interesting but, apparently, the authors do not analyze the basic mechanism behind the functioning of the DCM. At a first order of approximation the DCM is the depth where the upward diffusive nutrient flux is fully uptaken by, prevalently, phytoplankton. This is why the isolume is a good, first order, proxy for DCM depth. This has been discussed by Letelier e al. (L&O 49(2), 508-519, 2004) and, more recently by Cullen (AnnRevMarSci, 2015) none of whom is cited in the paper. Of course, there might be phylogenetic or ontogenetic adaptations, but I assume that the model has a constant physiology for phytoplankton. I would hypothesize that increasing vertical diffusion should certainly increase the carrying capacity, and therefore, the DCM amplitude, which is what the authors observe, but it should also move the DCM depth upwards to reach a new steady state where the nutrient flux is utilized at the higher rate because of higher photon flux. Diffusion does also disperse cells but the authors focus mostly on this aspect, i.e., the thickness of the DCM not on its depth dependence on diffusion and vertical gradients.
3. The authors focus on phosphate as the possible driving nutrient. It may depend on the existing paradigm that phosphate is the ‘limiting’ nutrient in the Mediterranean sea. It might be interesting to examine the nitrate behavior. However the intriguing pattern is that the phosphate concentration in the WMED is approximately double than in the EMED at the same isolume but the chlorophyll is more or less the same, for what can be seen from figures 3 to 6. How the authors interpret this, since the phytoplankton physiology should be the same? May be that the similarity is a bias of the graphic representation.
4. Linked to the above is the sensitivity of the DCM depth to phosphate (nutrient?) profile. Swapping East and West Argo profilers the authors (see response) state that there is no significant effect. Indeed, the slope of the model DCM vs observed DCM depth shows that the model underestimates the DCM depth for deep DCM and slightly overestimates the depth for shallow DCM, a pattern that is not discussed. More important, when they swap the profiles the model enhances this feature, which I would interpret as the fact that the higher irradiance in the EMED produce shallower DCM than in the real environment and the opposite occurs in the WMED. The scatter plot in Figure 2 does not allow a simple geo-localization of the Argos but, in any case, I would not consider the result of the analysis as a demonstration that nutrient profile has a minor role in determining the DCM depth.
5. More important, even not being a English mother language, I think that the text should be revised both in the wording and in the way the work done is presented. I still found some parts hard to follow and to connect to the others.