hessd-2022-166 “Does non-stationarity induced by multiyear drought invalidate the paired catchment method?” Y.Zhang, L.Cheng, L.Zhang, S.Qin, L.Liu, P.Liu, Y.Liu

This article looks at methods for partitioning changes in the rainfall-runoff relationship between vegetation changes, climatic variation, and non-stationarity in runoff generation.

The three acronyms PCM, TTM and SBM are introduced and explained in parentheses by the second sentence of the Abstract, but not until the third paragraph of the Introduction and then after the acronyms are already used. Please use and explain them as soon as they are mentioned in the main text of the paper.

Is the Pettitt (1975) method used in this work? It is mentioned once (p4 l101) where the authors state that the Mann-Kendall test is used for ranking tests of non-stationarity then never again.

Question of equilibration between rainfall-runoff process within catchment and between paired catchments? The catchments must be small enough and the changes of a suitable scale that either both equilibrate quickly, or at the same rate so that cumulative fluxes still appear as a straight line.

Do the authors think that having a single effect occur with reasonable gaps is necessary to analyse the data? In the case of Red Hill/Kileys Run the catchments were paired (hydrologically) well then had the afforestation and years of data, then the drought with years of data, then the post-drought conditions and again with years of data. Is there a risk if changing climate conditions inducing a non-stationary response would interfere with the land-use response if they occurred closely chronologically? Could a method determine the changes and separate them?

 General comments

1. In general it’s clear that the authors improved the original manuscript (https://doi.org/10.5194/hess-2021-5). They did improved the flow of the storyline and the analysis of the control catchment. The manuscript has appropriate objectives and gives insight in the change in hydrological processes caused by the combination of land use and climate changes.
2. Although the assumptions and fallowing conclusions are constructively fine and according to the applied methods, the arguments (especially in the abstract and conclusion) are not very clear. Although some comparative values per method are presented, for a reader it is not clear where the conclusions are based on.

Specific comments

3. The abstract ends with “paired-catchment method is proven to be still the moist reliable method even the control catchment experienced climate-induced shift in rainfall-runoff relationship”, but within the abstract no information on the control catchments are presented. I would suggest to give actual numbers so that the reader is able to draw this conclusion himself.
4. In my opinion it’s a missed opportunity that the majority of input by the previous reviewers and especially the replies by the authors are not processed within this manuscript.
   10. e. reply to reviewer 1 (https://doi.org/10.5194/hess-2021-5-AC1): “During 1988 the uppermost 50 ha of Red Hill was planted to p. radiata, with the remaining 145 ha planted to p. radiata in April 1989. During 2003 the plantations in Red Hill catchment were thinned to remove pulpwood and to promote the growth of sawlogs in the remaining stands. During the prolonged drought period, no trees died in the treated catchment. The change of annual PET can be seen in Figure 8 (b) (Page 42). PET showed an insignificant (p-value>0.1) increasing trend of 3.5 mm year-1. PET initially decreased before 1996 and then increased after 1996. The mean annual PET during the period of 1990-1996, 1997-2009, 2010-2015 and 1997-2015 are 1168 mm, 1262 mm, 1186 mm and 1238 mm, respectively. Compared with the period of 1990-1996 and 2010-2015, the mean annual PET during the period of 1997-2009 (the prolonged drought occurred) increased by 94 mm and 76 mm, respectively. Compared with the period of 1990-1996, the mean annual PET during the period of 1997-2015 increased by 70 mm. It is consistent with the cognition that afforestation and drought can make PET increase.”
   11. e. reviewer 2 (https://doi.org/10.5194/hess-2021-5-AC2), which you may use in the introduction and/or discussion: “Considering the influence of prolonged drought on the rainfall-runoff relationship in the treated catchment, the result of the paired catchment method is closest to the real runoff change caused by vegetation change. Because the control catchment indirect eliminate the influence of prolonged drought and climate variability on the treated catchment under the assumption that the response of the two catchments to prolonged drought is similar. Interannual changes in watershed storage occur primarily in soil water and shallow groundwater, pools that are often hydrologically active at time scales shorter than 1 year (Sayama et al., 2011). Rice and Emanuel (2019) indicates that down‐regulation of transpiration and inhibition of hydrologic connectivity by forest vegetation represent two important negative feedback processes that can avert large losses in soil water or plant‐accessible groundwater during dry periods. In doing so, this feedback mechanism has the potential to reinforce steady‐state (or near‐steady‐state) conditions in dry conditions. So, the sensitivity‐based method may be less affected by the prolonged drought because it is used in the forest and the time scale of PET and P data used in this method is annual scale. Runoff changes calculated by the sensitivity‐based method are induced by climate variability.”
5. For some statements I suggest to add some additional more recent papers.
   33. e. at line 33.
   34. e. at lines 52-53 “in many catchments around the world” two papers are cited one from 1980 and one from 2005, may be implying that these methods are not used that often anymore?
6. Lines 79-93: The data and location description are very brief. I suggest to give some more extended information. i.e. Ït can be seen that Kileys Run experienced a multiyear drought that lasted … with the period of the Millennium Drought”, explain why? What is lowest amount of rainfall measured?
7. Lines 346-357, the paragraph about “Multiyear drought induced changes in the rainfall-runoff relationship” is short and doesn’t discuss the subject. I suggest to compare your results with other locations where multiyear drought led to changes in the rainfall-runoff relationship.
8. In this paragraph you do mention a change in physical processes (runoff, soil moisture and evapotranspiration?), but not really compared with studies elsewhere. Do you have any specific evidence available about changes processes?
9. In addition to that, add some references about global and local knowledge about land use changes and their effects, such as the infiltration trade-off hypothesis (i.e. Bruijnzeel, 1989) and regional water availability caused by tree restoration such as (i.e. Hoek van Dijke, et al. 2022).
10. Lines 412-426, I suggest to re-introduce abbreviations.
11. Lines 417-418, conclusion. Add numbers so the reader is able to “agree” with your conclusion.
12. Idem line 420.

Technical corrections

13. Line 82, what about Red Hill?
14. Line 87, it would be nice to apply your analysis with data up to 2020?
15. Line 87, which method did you used to measure the runoff? Of where did you collected the data?
16. Lines 88-89, add more information/background. You use daily rainfall and runoff, but monthly PET?
17. Line 90, “figure 2 shows the rainfall anomaly that was calculated by the method proposed by”, I suggest to describe the method (define anomaly). Which in this case is the annual percentage of rainfall being larger or less than the average rainfall.
18. Line 91-92, I suggest add to add the reference again.
19. Figures 3 – 5, and 7: used colours (combination of red and green) are not very suitable for colour blind readers (HESS - Submission (hydrology-and-earth-system-sciences.net))
20. Lines 363-365, add reference?
21. Line 372, add reference?
22. Figure 4, perhaps add cumulative rainfall and the separated periods.

This is an interesting paper on an important topic. In fact it could be argued the issues discussed in this paper are absolutely fundamental in hydrology. This is particularly true in light of the way “data” is used uncritically in many studies.

The paper is generally well written and structured, although there is a need for a more careful read-through as there is some awkward syntax and grammar (particularly in the Discussion) where the quality of the writing seems to wander a little. My main questions relate to the data set used and the essential paradigm of the PCM. The Redhill site did not have a calibration period. Although the authors suggest that the first period of treatment may be thought of as non-treated in that the trees were very small and not high water using, I feel the implications of this may be important.  This then connects to the PCM paradigm; that is, that the length of calibration or the approach developing the calibrations in theory should account for the type of non-stationarity that is discussed (ie. drought). Putting this another way, how can we decide what is non-stationarity and what is variability? This is particularly germane to Australian hydrology where we experience significant variability. I am not suggesting climates are stationary, but disentangling non-stationarity from variability with a relatively short data period (in climate terms) is a question.

The calibration period issue is a vexed one as there is no longer an appetite by funding bodies to set up a paired-catchment experiment and then wait for a lengthy period before anything happens. Bren and Lane (2014, JH 519) explored this issue and proposed a method using daily flows that rather obviously increases the number of data points. Somewhat surprisingly the analysis showed that good calibrations (Nash-Sutcliffe E = 0.8) with 100 days of data, and very little improvement after 3 years. Apologies for the treatise, but I wonder if this approach might be useful in thinking about the PCM. That is, if such an analysis was performed and compared with the other analyses it might be very useful. Data could be randomly pulled out of the Kileys Run data. At the very least it should be discussed.

I have some more specific comments as follows:

Line 40 - there are more updated references for the research in Australian catchment behaviour that are relevant (eg Petersen et al, 2021,  Science 372)

L46. I think this statement about PCM and non-stationarity requires more justification. How does it not deal with the issue given that is the paradigm of PCM?

L 65+ Both TTM and SBM require lengthy records; is this an issue with these analyses at Redhill? There may be an argument to see the drought as a plus in terms of a record with wet/dry periods.

Lane et al. 2005 (JH) used FDCs that included Redhill – might be worth including these results as a comparison from a different method. This paper also has some estimates of time to equilibrium.

L 69 – syntax not great “issues about this hypothesis” could be improved.

L 107 – should be double mass curves, FDCs etc. There are quite a few examples of this, need a careful read.

L 168- See earlier general discussion. It does trouble me that a site with no calibration is used for this study. In addition, the vegetation effect is dynamic; growing from seedlings to (presumably, given there are no growth data) a closed canopy. I do wonder if this really is the best data set for such a study, or what might be gained from using more data sets.

Figure 3 is a great figure!

Table 2 – total flow changes would be useful. They appear later in the text but I think having totals in the table make it easier to evaluate the methods.

This also brings up another point that I don’t think has been discussed properly. Qclim is conceptualised as the climate effect, encompassing wet and dry and mean climate inputs. I am not sure there is adequate discussion of how this does not deal with the climate issue as formulated.

5.2.1 – this paragraph brings up the interesting point (that is the subject of the Saft/Peterson/Fowler etc studies); is it the climate that is non-stationary or is the processes (obviously driven by the climate).

The Paragraph around Line 375 needs some rewriting, the syntax is jarring. For example “the” control..

L 388 “Because Saft..” this is a poor sentence

L 399 “pines” should not be italicised P.radiata would be