Precipitation and Streamflow Reconstructions from Tree Rings for the Lower Kızılırmak River Basin, Turkey

Round 1

Reviewer 1 Report

Reviewer comments for Manuscript forests-1609816

“Precipitation and streamflow reconstructions from tree rings for the lower Kızılırmak River Basin, Turkey”

Specific Comments:

Abstract:

 

Somewhere in the abstract, recommend mentioning the span of the reconstructions, e.g. 1794-2003. I know you have enough info to figure it out, but it’ll make it easier on the reader.

 

L35: tree-ring studies

L65: the KRB

 

Reconstruction method: First, I want to say that your reconstruction methods are sound. I want to be clear that I am not suggesting that you change your reconstruction method…as presented, your findings are sound and publishable. However, if you applied a nested PC approach (e.g. Meko, 1997; Cook et al., 1999, 2002), you could add over 100 years to your reconstructions by being able to include info from the HAC chronology that extends beyond 1794. Given the EPS of HAC goes to 1689, you could extend your reocns from 1794 to 1689. I’ve also used this method for all reconstructions I’ve done (e.g. Harley et al. 2017 https://doi.org/10.1016/j.jhydrol.2016.11.020, so please feel free to contact me [email protected] if you have any questions about doing this yourself).

 

Meko, D.M., 1997. Dendroclimatic reconstruction with time varying predictor subset of tree indices, J Clim, 10, 687–696.

Cook, E., Meko, D.M., Stahle, D.W., Cleaveland, M., 1999. Drought reconstructions for the continental United States, J Clim, 12, 1145–1162.

Cook, E., D’Arrigo, R., Mann, M., 2002. A well-verified, multiproxy reconstruction of the winter North Atlantic Oscillation index since A.D. 1400, J Clim, 15, 1754–1764.

 

Speaking of reconstruction methods, I’m unclear of which exact chronologies were pulled into each recon model, e.g, why wasn’t KAR used? And if it was used, what is the EPS of that reconstruction? The only line that alludes to this is L159. I’m guessing that KAR extends well beyond 1689 (when HAC ends), so again, if you used a nested approach, you could provide much longer reconstructions of precip and sreamflow.

 

Figure 4: I like this figure, but I think it needs some work. First, the text in the fig is very small and once published, there’s no way it will be readable. Here’s what I would do (the easiest way to improve it, in my opinion) …panel (b) is twice the size as it needs to be, plus, there’s a bunch of white space in panel A. I would just show the corr matrix, given that all coefficients are sig @ p<0.001 (except MEM and IKI, but you could just say that in the caption). Move the corr matrix and put it in the white space of panel A.

 

Figure 6: the light gray is hard to see.

 

Figure 7: The two recons look very similar, but this isn’t a huge surprise given the climatology of the region. I think you need to expand more on how and why both recons are highly correlated. You also need to be clear about why it’s important/needed for you to produce two reconstructions for two variables that are basically the same. How correlated are the 2 instrumental datasets? I just think it’s odd to include both a precip and streamflow recon that use the same predictor chronologies in the same paper. Now, if you were to discuss the differences in the recons (residuals) and mention how snowpack inputs relate to this, that could be a way of discussion this issue.

 

In an area where summer precip and streamflow are likely highly correlated, can you explain the major purpose of producing 2 reconstructions using the same input chronologies (all with climate signals driven by moisture availability)? If the 2 recons are the same, are they really adding much information?

Author Response

Response to Reviewer 1 comments

We thank for your valuable comments and suggestions, which improved our manuscript. We listed our response to your comments point by point below:

Suggestion: Somewhere in the abstract, recommend mentioning the span of the reconstructions, e.g. 1794-2003. I know you have enough info to figure it out, but it’ll make it easier on the reader.

Answer: We added time-span of the manuscript (1794-2003) to Abstract.

Suggestion: L35: tree-ring studies

Answer: Corrected

Suggestion: L65: the KRB

Answer: Corrected

Suggestion: Reconstruction method: First, I want to say that your reconstruction methods are sound. I want to be clear that I am not suggesting that you change your reconstruction method…as presented, your findings are sound and publishable. However, if you applied a nested PC approach (e.g. Meko, 1997; Cook et al., 1999, 2002), you could add over 100 years to your reconstructions by being able to include info from the HAC chronology that extends beyond 1794. Given the EPS of HAC goes to 1689, you could extend your reocns from 1794 to 1689. I’ve also used this method for all reconstructions I’ve done (e.g. Harley et al. 2017 https://doi.org/10.1016/j.jhydrol.2016.11.020, so please feel free to contact me [email protected] if you have any questions about doing this yourself).

Answer: Thank you for your suggestion. We’ve already applied the nested approach to go back earlier. But our model which uses less chronologies, did not give statistically significant verification values. Therefore, we could not use nested approach and presented the model which only uses all chronologies.

Suggestion: Speaking of reconstruction methods, I’m unclear of which exact chronologies were pulled into each recon model, e.g, why wasn’t KAR used? And if it was used, what is the EPS of that reconstruction? The only line that alludes to this is L159. I’m guessing that KAR extends well beyond 1689 (when HAC ends), so again, if you used a nested approach, you could provide much longer reconstructions of precip and sreamflow.

Answer: We’ve used KAR chronology too in both current reconstruction and tried nested approach. We’ve clearly gave the chronology codes of previous studies in L166. As we mentioned previously nested approach did not provide significant verification values, which means did not pass verification tests. In the method section in “Hydroclimate reconstructions” we have already mentioned that we used first two PCs of seven chronologies for both reconstructions.

Suggestion: Figure 4: I like this figure, but I think it needs some work. First, the text in the fig is very small and once published, there’s no way it will be readable. Here’s what I would do (the easiest way to improve it, in my opinion) …panel (b) is twice the size as it needs to be, plus, there’s a bunch of white space in panel A. I would just show the corr matrix, given that all coefficients are sig @ p<0.001 (except MEM and IKI, but you could just say that in the caption). Move the corr matrix and put it in the white space of panel A.

Answer: We re-drawn the figure and changed the caption as you suggested.

Suggestion: Figure 6: the light gray is hard to see.

Answer: We re-drawn the figure as you suggested.

Suggestions:

• Figure 7: The two recons look very similar, but this isn’t a huge surprise given the climatology of the region. I think you need to expand more on how and why both recons are highly correlated. You also need to be clear about why it’s important/needed for you to produce two reconstructions for two variables that are basically the same. How correlated are the 2 instrumental datasets? I just think it’s odd to include both a precip and streamflow recon that use the same predictor chronologies in the same paper. Now, if you were to discuss the differences in the recons (residuals) and mention how snowpack inputs relate to this, that could be a way of discussion this issue.
• In an area where summer precip and streamflow are likely highly correlated, can you explain the major purpose of producing 2 reconstructions using the same input chronologies (all with climate signals driven by moisture availability)? If the 2 recons are the same, are they really adding much information?

Answer: The reason why we reconstruct very highly correlated precipitation and streamflow data is that our limitation due to quite short instrumental streamflow data (45 year-long, 1954-1998) to apply a split-calibration method. Therefore, we also provided precipitation reconstruction and used it to verify our streamflow reconstruction. We’ve already explained that in the last paragraph of “Method”s in “Hydroclimate reconstructions”).

Reviewer 2 Report

General remarks of the reviewer

Title: The title of the article is accurate and directly relates to the purpose of the research.

Abstract: The abstract gives a good overview of the work.

Keywords: The keywords are specific to the topic under study.

Introduction: The state of the research  is reviewed  and major publications cited.

At the end of the introduction, please clearly articulate the research goal and research hypotheses.

Materials and Methods:

2.3. Streamflow and Climate Data

In the legend of Figure 3a, please add an explanation of the descriptive statistics, and for Figure 3b, include the trend functions along with the determination coefficients.

Results:

The results were correctly statistically analyzed and well presented graphically and in tables.

 

Discussion: The research results are well-discussed.

Conclusions: On the basis of the research carried out, please provide a few specific conclusions of a cognitive and utilitarian nature.

References: For some newer publications, please add DOI.

Technical Notes

Please correct any minor inaccuracies in the text of the article.

The description of the literature item needs to be corrected as required by the publisher: articles, books and other sources - italics of journal titles, year in bold, correct pages of journals and the access link and date of access in English. According to MDPI standard.

Details in the attached manuscript.

Summary of the review:

The article exhausts the presented issue.

Comments for author File: Comments.pdf

Author Response

Response to Reviewer 2 comments

We thank for your valuable comments and suggestions, which improved our manuscript. We listed our response to your comments point by point below:

Suggestion: At the end of the introduction, please clearly articulate the research goal and research hypotheses.

Answer: We added a clear research goal and hypotheses at the end of introduction.

Suggestion: In the legend of Figure 3a, please add an explanation of the descriptive statistics, and for Figure 3b, include the trend functions along with the determination coefficients.

Answer: We added an explanation of the descriptive statistics in the caption of Fig. 3a, and included the trend functions in Fig. 3b.

Suggestion: On the basis of the research carried out, please provide a few specific conclusions of a cognitive and utilitarian nature.

Answer: We provided a few specific conclusions.

Suggestion: Please correct any minor inaccuracies in the text of the article.

Answer: We corrected.

Suggestion: The description of the literature item needs to be corrected as required by the publisher: articles, books and other sources - italics of journal titles, year in bold, correct pages of journals and the access link and date of access in English. According to MDPI standard.

Answer: We rearranged the cited literature as you suggested.