An Empirical Research on Influence Factors of Industrial Water Use
Round 1
Reviewer 1 Report
The paper reports on how Vector autoregression model is applied to study the dynamic influences of industrial development, technological progress and environmental protection on industrial water use change in Jiangsu Province, China from 2001-2015. The presents an interesting study, which merits publication in the journal. However, I recommend the author to reflect upon the following aspects:
Introduction: over the last decades, multi-purpose water management reflects the different functions provided by rivers, which can be complementary or conflicting see Van der Voorn et al (2018). This calls for more integrated water resources management. In the Discussion section, I would like to see a more critical reflection on how the Vector autoregression model can support IWRM Results: how do the authors account for model uncertainties? Conclusion: the authors acknowledge the limitations of the study. I would like to see a more critical reflection on these limitations in the Discussion section: How to account for the size of data sample? How to account for the uncertainty about the relationship between CC effects on industrial water change?
References
Van der Voorn, T.; Quist, J. Analysing the Role of Visions, Agency, and Niches in Historical Transitions in Watershed Management in the Lower Mississippi River. Water 2018, 10, 1845.
Author Response
Response to Reviewer 1 Comments
Point 1: Introduction: over the last decades, multi-purpose water management reflects the different functions provided by rivers, which can be complementary or conflicting see Van der Voorn et al (2018). This calls for more integrated water resources management. 

Response 1: this is a very good references for this research, and the content has been added to the Introduction.
Point 2: Results: how do the authors account for model uncertainties?
Response 2: The VAR model uncertainties include input data of model, parameterization of model, and model structure. To reduce model uncertainties, unit root test, optional lag order selection and model stability test were applied. The unit root test ensures the stability of time series of model variables; the choice of the optimal lag order makes it possible to make full use of the variable information of the constructed model; model stability test is the premise of impulse response analysis. Therefore, section 3.2, 3.3, and 3.4 reflected the source of model uncertainty, and utilized appropriate methods and criteria to effectively reduce the uncertainty of model, and ensure the scientific and reasonable results of the model.
Point 3: In the Discussion section, I would like to see a more critical reflection on how the Vector autoregression model can support IWRM. I would like to see a more critical reflection on these limitations in the Discussion section: How to account for the size of data sample? How to account for the uncertainty about the relationship between CC effects on industrial water change?
Response 3: In sum, the VAR model could quantitatively reflect the degree of influence of industrial development, technological progress and environmental protection on the change of industrial water use, which is conducive to the authorities to propose targeted integrated water resources management measures to solve the problems faced by industrial water use caused by external environmental change. Meanwhile, it reveals dynamic relationship among industrial water use and industrial development, technological progress, and environmental protection, which facilitates policy makers to timely adjust integrated water resources management measures and promote dynamic water resources management.
Furthermore, the critical reflection on the limitations of this research should also be discussed. One is the size of data sample. Due to the lack of statistics, only 15 data samples are exposed for each variable. However, the time series passed the stationarity and model stability test, which indicates that the process of model construction is reasonable and the output is credible. Of course, if the number of samples of the variable is large enough, the relationship between variables could be reflected more comprehensively. The other aspect is the selection of variables. We aim to clarify the impact of economic and social development on industrial water use in this paper, that’s why we neglect to analysis the influence of climate change on industrial water use. In fact, climate change has direct and indirect effects on industrial water use [54, 55]. On the one hand, cooling water is a major part of industry water use. Climate change causes temperature change, and temperature change directly affects cooling water consumption. On the other hand, industry is an important part of carbon emissions. Adaptation and mitigation measures of climate change will affect the development scale of industrial industry and indirectly affect industrial water use. The impact of climate change on industrial water use is complex and uncertain, it needs to conduct in-depth research based on the results of this work.
Reviewer 2 Report
A well written paper assessing influential factors on industrial water use.
The introduction presents the topic adequately and also includes a brief literature review, in order to justify the novelty and objectives of the paper in the last paragraph.
My main concern is related to the methodology. I do not know how the authors can know how the model is going to result aprioristically, which is why they propose the transformation expressed in Eq. (2). The model should have been tested without any modification, such that the results obtained from applying this original model led to identify the most suitable transformation to maximise the fit. Furthermore, there are multiple measures that account for heteroscedasticity and multicollinearity without requiring any transformation.
The results are properly presented and discussed, including relevant tables, figures and references to support them. Finally, the conclusions are well approached. They deal with the implications derived from the research, while highlighting its limitations and how they could be amended in future investigations.
Author Response
Response to Reviewer 2 Comments
Point 1: My main concern is related to the methodology. I do not know how the authors can know how the model is going to result aprioristically, which is why they propose the transformation expressed in Eq. (2). The model should have been tested without any modification, such that the results obtained from applying this original model led to identify the most suitable transformation to maximize the fit. Furthermore, there are multiple measures that account for heteroscedasticity and multicollinearity without requiring any transformation.
Response 1: The mathematical expressions of VAR (p) model can be depicted as Yt =u+A1·Yt-1 + … +Ap·Yt-p+et (t=1, 2, … , T) (Eq. (1)), and it is a general standard reduced-form. In this work, the VAR system is constructed using the following four variables: industrial water use (Wt), industrial development (Gt), technological progress (It), and environmental protection (Et). Eq. (2) is the embodiment of specific form of Eq. (1), and no modifications have been made to the original model.
In this paper, the dimensions of the four variables are different, and there are orders of magnitude differences. Logarithmic treatment is an effective method, which could avoid the sharp fluctuation of data and eliminate the possible heteroscedasticity, and not change the characteristics of the time series data. In addition, logarithmic treatment was commonly used in VAR model variables time series stationarity test in previous studies. Therefore, the transformation of time series data is beneficial to the application the VAR model.
Point 2: The results are properly presented and discussed, including relevant tables, figures and references to support them. Finally, the conclusions are well approached. They deal with the implications derived from the research, while highlighting its limitations and how they could be amended in future investigations.
Response 2: thank you very much for the valuable suggestion, we improved it again and hope this will be suitable for publican.
