The authors want to use the Landlab landscape evolution model coupling an evapotranspiration model to investigate the direct and indirect effects of anthropogenic land cover change on hydrological and morphological processes in the Wei River catchment since the mid-Holocene. However, two model scenarios (a model with land use and climate change, Normal, and a model without climate change, WCC) used in the Holocene simulations were equipped to show the impacts of climate change rather than land use change, the importance of which is though shown in Discussion. The scientific question mentioned in Introduction does not been directly answer. Lack of necessary information of the methods and data makes readers difficult to understand fully only by their paper. Many uncertainties originated in paleoclimate data, model parameters, model resolution, and initial topography in 6000 BCE make the results less convincible. Thus, I do not recommend the publication of current version in the journal of ESD. 

1.    The numerical modeling is the core method. However, it is not introduced enough in Introduction. Applications of landscape evolution modeling for other similar studies should be mentioned. 

2.    Line 85: what is “KK10 scenarios” in this sentence? We do not know anything about KK10 before the line. You should briefly describe what is KK10.

3.    Line 99: the study area is not an East Asian monsoon region but just belongs to it.

4.    Line 157: what is the time scale? daily, monthly, or yearly? what is the time range of observed data? 

5.    Line 183-198: a technology roadmap of model development should be given here.

6.    Line 189: please describe what the tuning method looks like.

7.    Line 225: in the Fig S1, sample numbers for validation may be too small, which means the high R2 cannot pass the significance test. It may bring more uncertainties into the following results.

8.    Line 226-229: “the reconstructions of Holocene climate including precipitation and temperature (Peterse et al., 2011; Chen et al., 2015) are used to predict the climatic inputs, by methods of Chen et al. (2001).” makes me confused. Did you use Holocene precipitation and temperature to predict climatic inputs? What do climatic inputs mean here? What is the method of Chen et al. (2001)? In addition, the word “predicted” is inappropriate, it may be some words like “simulated” because it is nothing about the future. 

9.    Line 238: how about the uncertainty of KK10 database calculated from a global ALCC model for the local catchment in this study?

10.    Line255-260: how did authors guarantee the reasonability of initial topography? If not, the study may be inauthentic and become a sensitivity test.

11.    Line 270: what is the method proposed by Chang et al.(2016)?

12.    Line 272-273: please show the acknowledged evaluation criterion and its results in a figure for the mentioned accepted calibration results.

13.    Line 315: what are the climatic inputs for Normal and WCC scenarios, respectively? What is the temporal and spatial resolution of model outputs? 

14.    Line 345: why constant for each 500 years? 

15.    Line 354: what is the relation between runoff and discharge? how did authors calculate the discharge by runoff?

16.    From the difference of model results for the two scenarios, impacts of climate change can be identified but no evidence shows the effects of ALCC. 

17.    In section 5.1, the sensitivity of discharge and sediment yield to the climate changes is based on statistical analysis rather than the ALCC impact on land-air interaction. It belongs to a complex climate system, so the title “Regime shift of a large river as a response to Holocene climate change depends on land use” is difficult to be addressed mechanically without climate-landscape evolution coupled modeling. 

The manuscript of Regime shift of a large river as a response to Holocene climate

change depends on land use – a numerical case study from the Chinese Loess Plateau presented a study that investigated the direct and indirect effects of ALCC on hydrological and morphological processes in the Wei River catchment since the mid-Holocene. After reading through the manuscript carefully, I think it’s an interesting study, but I still have several minor concerns. And I think the manuscript needs at least minor revision before be published in ESD. Here are some reasons why.

1.    Line 85: What is KK10？I think authors should add more information about it in the supporting information since it’s maybe one of reasons for the sources of sensitivity.

2.    Line 184: “model development” seems to be inaccurate in here, and I think “model summary” better sums up Sect 3.1.

3.    Line 271: Is there any reason why the authors think that a 10% error is acceptable? In my image it is usually 5%.

This paper aims to understand the relationship between human activity, climate change, sediment yield, and discharge in the Wei River, a tributary to the Yellow River. The authors use a series of nested models to do this analysis. The sediment and discharge models are calibrated to modern conditions using a robust record from Chinese gauging stations. They then run simulations for past climate and land use scenarios using previously published modeled data for these conditions. Overall, I found the paper interesting and the story compelling. Although it is not surprising that we find increased sedimentation when human activity increases, it is interesting to consider the interplay between anthropogenic activity and climate change in driving the changes we know happened in the past. The biggest concern I have is that the authors do not connect their work very explicitly to other work done on human-land use and erosion in China. 

I am not an expert in numeric models and so this review is largely from a geomorphology perspective rather than on the model simulations themselves. I do have a few questions or points for the authors to clarify though.

• It isn’t clear to me why some decisions for model parameters were made based on totally modern agricultural practices. The model requires soil nitrogen content and this was set as constant based on modern fertilization levels. I don’t understand why this amount was chosen when fertilization would have varied over history. Similarly, the crop is assumed to be irrigated twice, but irrigation would have been different at different points in history. Finally, why was winter wheat chosen for all time steps? Do we know that that is what people grew in the past? If not, do we have data that could provide better information for past crop growth?
• I had a hard time following all the different geographic names, locations, watersheds, gauging stations, and so on. I know these data are available in figure 1, but it is such important information and so hard to read in such a small figure. It might help readers who are less intimately familiar with the area to have a larger context map with the rivers, gauging stations, and sub-catchments clearly delineated.
• Several times in the results the authors have long lists of results that would be better presented in tables. It’s very hard to follow long lists of results, especially for people who are less familiar with the study area.
• I’m concerned that the humidity is set to modern levels but we know that various times in the Holocene had very different humidity levels based on loess vs soil accumulation on the Chinese Loess Plateau. I wonder if this would then affect vegetation and erosion.
• Why was a 90 m DEM resampled (with, I assume, interpolation, although the exact interpolation was not specified) when there are high quality 30 m DEMs now available? It seems like an additional level of uncertainty that could have been avoided.
• The authors talk a lot about base layers and surface layers but never specify what they are using for base and surface layers. Is the base layer loess? If so, how is loess thickness set when loess is continuously being deposited? Regardless, it would be good to explicitly say what the base and surface layers are composed of.
• What is the correction used for dams and irrigation channels? How do we know that is reasonable in the past?
• All the web resources the authors list (mostly data sources) should be properly cited rather than just webpages listed in the text.
• In the discussion, I wasn’t convinced by the argument that climate change is less important than human activity. It needs more explanation.
• Likewise, I didn’t understand the argument about the multiple thresholds of sensitivity to change. This seems interesting but was not well enough explained.
• It would be interesting to see this work connected to other human-land use work in China. There is such a rich literature on this topic, including from lake cores, simulations, and geochemistry. This paper would have a much bigger impact if it drew those connections more explicitly.
• There are a few typos in figures and tables that need to be fixed. For example, table 1 has ages for Yangshao that don’t overlap the time period for that row. Figure 5c has a typo in the title. In general, figures that have small text and a mix of red to green colors can be hard for many people to read.

I think a lot of my concerns are mainly just more explicitly connecting dots for readers and making it easier for readers to see how to get through various steps that the authors are more intimately familiar with. Once all this is fixed up, I do think it is a really interesting paper with neat results.

Amanda Schmidt

Oberlin OH, USA