Using a process-based LPJ-GUESS model, this paper investigated the linkage between the dynamics of terrestrial nitrogen in the Yangtze Plain and eutrophication changes in fifty large lakes for the past four decades. Overall, the paper is well organized and the results answer the questions mentioned by the authors, although some uncertainties exist. But some details regarding the innovation and the method need to add before publication. The specific issues are as follows:

1 Introduction 

Suggest to strengthen the innovation of the study.  

2 Materials and Methods

2.1 briefly introduce the agricultural cultivation types and management patterns in the study area.

2.2 How the fertilization process is represented in the model needs to be described

2.5.1 The remote sensing data of Guan, such as spatial and temporal resolution, should be briefly introduced.

4 Discussions and conclusions 

 (1) Lines 343-348 add a discussion of the reason for the decrease in NUE. 

 (2) add an outlook for future studies regarding the uncertainties of current results.

Comments on the study of Guan et al

This study examined the dynamics of major pollution sources contributing to eutrophication in the fifty large lakes of the Yangtze River Plain, a region with 340 million people. These dynamics were studied over a forty-year period (1979-2018) and correlated with satellite-tracked algal bloom events. Pollutant sources studied include leached nitrogen in agro-ecosystems, phosphorus sources, and industrial waste. Remarkable efforts have been made to characterize in space and time the leaching of nitrogen which is a very diffuse source of pollution. This characterization was made using a carbon-nitrogen coupled ecosystem model constrained with meteorological, soil and agronomic data. This study is exceptional by the size of the studied region, the diversity of ecosystems considered (crops, terrestrial and aquatic ecosystems) and used variables. Below I made some comments/criticisms with the aim of improving this manuscript that merit publication.

The title and conclusive sentences of the abstract only mention the impact of N leaching. However, your study also suggests an impact of P applications and industrial wastes. This must appear more clearly.

From my point of view, on the most important finding of your study is the suggestion that the origins of lake eutrophication change with space (and maybe with time). Indeed, this finding signifies that policies and technical changes that need to be implemented to decrease lake eutrophication must be adapted to local conditions.

In the discussion, you mention that the eutrophication induced by industrial wastes is now solved thanks to policies set by the government and changes made in industries. YOU CAN NOT CONCLUDE on that point because your data do not show that at all. You can only mention the fact that the government has become aware of the problem and set up incentives to decrease industrial pollutions.  

The manuscript discussion starts by saying that it is now possible to deplete lake eutrophication because we know one of its main sources: low NEU and high N leaching. One might argues that we know since decades that N leaching causes eutrophication. The challenge rather lies in the setup of good agricultural practices leading to a decrease in nutrient loss. You should mention/list some examples of these practices (agriculture precision, cover cropping, residues managements etc) and how they are implemented or not in this region. More generally, I would suggest to start the discussion with your finding that the causes of eutrophication can be multiple and change in space requiring incentive policies adapted to local conditions.

L153 It seems that soil properties have been implemented in the model using the world database on soil types/soil pedology. How did you consider the impact of land use on these soil properties?

L199 This approach using the linear relationship is not clear? Why did you not directly use the model to estimate the N leaching?

L212-213 I do not understand why it is a problem since the model can provide those results.

L217 P sources are not simulated right? This should be specified.

L225 These choices are a bit strange. In general, N (especially mineral N) is more mobile than P. Thus, N leaching is likely to be more linked to N applications than P leaching to P applications. Despite this, you used a complex modelling to estimate N leaching whereas you simply used the P source for P leaching/pollution. You should better explain your choices, and maybe consider to also examine the correlation between N leaching and N applications to have something comparable for the different sources of pollutions.

L256 What do you mean with climatological NEU?

L270-273 There are considerable differences between crops, which is worth to discuss. The negative effect of double crops is surprising since maintaining a permanent plant cover generally promote N retention in agrosystems.

Figure 6 -> N leaching under natural ecosystems also increase during the period, which merits discussions.

L292-293 -> put the treatments close to the numbers to facilitate the reading.

L298 How did you define the drainage area? Watershed?

L303 What about the role of atmospheric N depositions? Did you consider this aspect or not? This should be specified.

L305 -> please recall us the acronyms of these variable to facilitate the reading.

LL377 Revise the sentence. The source of uncertainty is not “potential impacts of terrestrial nutrient losses” but the transport processes that mediate the quantity and quality of nutrients that will be transferred from sources to surface water (lakes?).