Dear authors,

Thank you very much for your detailed responses to the very helpful reviewer comments. As you have noted there are several major concerns that have been raised and that need to be addressed in detail in the revised version of the manuscript:

(1) As pointed by all reviewers, and I fully agree, a model with >100 parameters does need some careful considerations and efficient methods to effectively constrain the feasible parameter space. In principal I commend you for you approach to identify sensitive parameters and to only calibrate these parameters for each model component. However, i am not convinced that this will be sufficient to robustly identify parameter sets that provide plausible model internal dynamics. I am surprised that you did neither (1) follow a more consistent multi-objective and multi-criteria calibration strategy nor (2) apply some kind of regularization or stronger prior information (e.g. relational parameter or process constraints). Making use of several orthogonal (!!) objective functions has in the past proven very valuable to identify unfeasible parameter sets. Similarly calibration against multiple variables/signatures (for the hydro module e.g. flow, flow duration curve, autocorrelation function of flow, etc.) has recently received considerable attention. To move away from pure curve fitting, hydrology and related sciences need to start to move away from curve fitting towards more process consistency and efficient model selection techniques. I would therefore encourage you to dig a bit into the recent literature and amend you calibration strategy accordingly, so as to ensure an efficiently constrained model.

(2) One reviewer also pointed out that it may be over-ambitious to publish such a comprehensive model together with a case study in one paper. In principal I agree with that, as in the original manuscript the descriptions of the model but also of the calibration strategy remained rather superficial. If you prefer to keep the entire manuscript in one, that is fine, but in the sense of reproducibility and dissemination of your model I would strongly encourage you to provide more detailed model descriptions (incl. the full set of equations) in the supplementary material.

(3) The manuscript lacks an adequate discussion so far. How does the model compare to similar models? What are its limitations? What are its advantages? Where should future improvements be necessary?

(4) It is not clear what the scientific objective of the manuscript is. What are the hypothesis you are intending to test? Please explicitly state this at the end of the introduction section

(5) The manuscript (also the newer version attached to the reviewer responses) requires detailed proof reading by a native speaker. Grammar and spelling errors are plenty.

I would be glad to received a revised version of your manuscript if you find it feasible to adapt it according to the points given above and the rest of the reviewer comments.

Best regards,

Markus Hrachowitz

Dear authors,

the three reviewers were in principle happy with the revisions in your manuscript as most of their comments were adequately addressed. Two reviewers pointed out some minor points which are given in detail below and which i would encourage you to fully address before he manuscript can be accepted for publication:

reviewer #2:

The abstract has a good balance, with an introduction, aims, methods, results and discussion/conclusion. The results section in the abstract is perhaps too long. It is confusing as to why a present tense is used in the abstract. I would prefer a past tense to describe results.

'Earth' should be spelt with a capital.
The manuscript mixes the use of a serial comma with non-use. A consistent usage or non-usage must be implemented. For example, at line 1 page 3.
Line 9 page 3 : 'Richard's'
Inconsistent citation formats at:
page 2 line 32
page 3 line 26
page 3 line 32
page 4 line 19
page 4 line 23
The use of a present tense here is also unusual and I don't feel comfortable with it as a reader.
Serial commas at:
Page 7 line 29
Page 8 line 6
Page 8 line 22
Page 9 line 4
Inconsistent citation format and serial comma at page 8 line 13
n-type dash to indicate a range: page 8 line 15
'Equifinality' is mentioned on page 11 line 15 but there is no attempt to explain what the term means or even provide a citation
Forests, water bodies etc. are referred to as land use, but are these not more representative of land cover?

Page 7 line 2-3: how is surface runoff different from overland flow?

Figure 4B: for the overland water quality module, are only two broad land use types considered, namely urban and rural areas? I would surely make sense to consider a greater range of land use types?

The WQM does not mention algal growth. Does the model simulate this and the effect on nutrients?

Not enough information is provided on the WQM in terms of simulating water quality in reservoirs: stratification? Water quality sink?

Page 10 line 11 ‘spatial consistent’. Please reformulate this sentence into readable English.

Page 15. In terms of ranges and using hyphens, use the n-type dash.

Page 15 line 3 and 4: incorrect use of brackets. Reformulate the sentence.

Page 15 line 30 to 32: use a thousands separator (,) as you have in other parts of the document.

Page 11 line 24: You mention the ‘NS’ to evaluate model performance. Please state what this is. I assume it is the Nash Sutcliffe efficiency? Please provide the citation.
Page 16 line 25: Using the log (NS) is less sensitive to extreme values.

I think the manuscript would benefit from some discussion on the strategy of more complex models (as taken here) as opposed to simpler models that attempt to simulate the most important processes explaining the majority of water quality variation (requisite simplicity), and the corresponding benefits/disadvantages of each broad type of model in regards to the required observed data, ease of use, and equifinality. Perhaps the model described here needs to be places in a more global context: could it be used in other countries? Would it be of benefit in developing countries where data and resources are a limitation?


reviewer #3:

Thank you for responding to most of my own and the other reviewers concerns. I have just a few minor points for revision.
1. scientific purpose - The authors state that the scientific purpose is the integration of multiple bio-geo-chemical proccesses. It would be nice if they could turn this into a question: "Does including these extra processes in an integrated manner improve model results compared to models that are focuses only on one component?" I think they show this to some extent in their comparisions with SWAT, but could slightly reword their introduction and discussion to better demonstrate this. The discussion now is a slightly boring comparison with SWAT. Remove the long list of number comparisions and discuss whether or not including more processes in the new model allows for an improvement of the overall model results compared to SWAT. Incidentally, if you really wanted to demonstrate the added value of integrated components, you could allow interation between the model modules in calibration, e.g. when calibrating water quality you may find issues to improve water quantity (such as the need for more surface runoff to increase particulate phosphorous erosion). This is why sometimes manual calibration can actually be superior to automatic.

2. Input Data: I'm not sure the authors understood my concerns about input data: My question to them is "Does the resolution and quality of the input data support the complexity of the process descriptions in the model?" How accurate are the agricultural N and P inputs or the point sources? Does the government have good control over this in China or do farmers and industries flaunt environmental regulations? THis varies considerably from country to country and therefore affects the quality of the model inputs. This was touched upon in the crop model results but should be more clearly mentioned in the discussion.

3. Some minor points:
- Please check the spelling and grammar one last time - there are some small errors remaining
- The HYPE model IS developed for integrated water quality and hydrology (N, P and to some extent C) and also includes regulation of dams. See Strömqvist et al. 2011.

Dear authors,

I am glad that you addressed in detail the remaining comments two of the reviewers. I think that this was done in an adequate way and that the manuscript can now be published as is. Congratulations on an interesting contribution to literature!