Dear Authors,
Please upload the revised manuscript according to the Referees' comments and your response, which will be subject to a re-review by Referees and editor. Thanks for the contribution.

Best,
Fuqiang

Please revise the manuscript accordingly based on the referee's report listed below.

The nonstationarity of the runoff in Wei River basin is very significant and this work applied multiple variables into time-varying model by GLM. The revised version addressed the comments of the last two referees clearly.

Line 26 and the others, “potential evapotranspiration, ET”. Usually, ET is used to represent actual evapotranspiration and EP is used to represent potential evapotranspiration. It’s better to use EP to represent potential evapotranspiration.

Irrigated area is a very important index in the Wei River due to large agricultural irrigation water withdrawn. And irrigated area is added in the revised version.

Line 345-348, “Human activity data(i.e. gross domestic product, population and irrigation area) were taken from annals of statistics provided by the Shaanxi Provincial Bureau of Statistics (http://www.shaanxitj.gov.cn/) and Gansu Provincial Bureau of Statistics (source: http://www.gstj.gov.cn/).” If the data also come from Zhang (2008) as shown in Line 326, it should be listed here. It should be noted that the “population” in the annals are different from the people who lives in the catchment. So the uncertainty should be presented here to remind the readers. Nonetheless, it is the best population data so far.

Dear Authors,
I am sorry to reply to your revisions later. I can see from your new manuscript that most reviewers’ comments are properly addresses. Yet, after my own another round of review, I think the follow aspects require the further revisions before the final publication:
1) The authors chose 3 human index and 8 climate/catchment factors as the potential explanatory variables for low flow decrease. The readers would like to know the rationality of such choosing. It would be perfect if the authors can add some discussions on this in section 3.2.
2) The authors conclude that the faster streamflow recession is one of most influential factors for low flow decrease. This is an interesting point. I suggest to add some discussions on the possibles causes of such increasing of streamflow recession, e.g. the human alterations on catchment properties or self-evolution driven by climate change and/or human interventions.
3) The authors cite quite a few papers in some sentences, an extreme case is in the first paragraph in page 5. I don’t think they are all relevant. I suggest to keep only the papers most relevant, which would be better for the readers to follow.
4) I think it is not necessity at all to discuss the Budyko type work here on Page 30.
5) In abstract, GA and AIC are not defined before. Please use them cautiously. Also, the statements about detail VIC values are not synoptic. I don’t think VIC values are popular to hydrology community. Please rewrite in a informative way.
I would like to re-review your further revised version as soon as possible. Wish you a pleasant and fruitful New Year 2018.

Best,
Fuqiang