Species Sensitivity to Hydrologic Whiplash in The Tree-Ring Record of the High Sierra Nevada
Round 1
Reviewer 1 Report
Dear Authors, Dear Editors,
The manuscript (Winitsky et al. Species Sensitivity to Hydrologic Whiplash in The Tree-Ring Record of the High Sierra Nevada) is an interesting contribution. It is well-known that tree-rings are valuable proxies of hydroclimate variability. The paper study an extreme mode of such hydroclimate changes. The results and the methodological improvements could be useful and informative also to other dendroclimatological researches.
The study is well-written and follows a logical structure. Beside the comments on Tab 3 and Fig7 all my other comments are minor or technical so, taking all together my opinion is that the study is publishable following a moderate revision.
Major comments on display materials:
Table 3: Please check the header of the table. Variable usage of uppercase and lowercase characters is a bit confusing.
Figure 7: Please check these plots. The legend seems to be incomplete and the axis labels are also missing.
specific comments:
line 49: Maybe “decreasing” instead of “decreased”
line 131: I think ITRDB might deserve a citation. e.g., https://doi.org/10.1177/0959683697007002
line 206: maybe m should be capitalized in the expression „M-or-more”
line 228 Please consider using conventional expression for powers of ten (10-5 ) instead of their digital equivalent 1E-5 (and the same comment for Table 3)
line 309: Perhaps pF>0.05 instead of F>0.05
line 403: What do you mean by “g species”?
line 503: Maybe spell out species name here to make the conclusions more understandable as a standalone text.
References
line 572: Could you update the bibliographical details? or at least remove “n/a(n/a).”
line 577: journal name is missing here
Author Response
Please see the attachment (uploaded pdf)
Author Response File: 
Author Response.pdf
Reviewer 2 Report
It can be published in this form.
Author Response
Reviewer #2 requested no changes
Reviewer 3 Report
Winitsky et al have compiled a large network of conifers from Sierra Nevada and explores if the tree-ring data is sensitive to so-called hydrologic whiplash. They find that there is a significant relationship of both positive and negative whiplash events for most species, but a weak and complex relationship on site-chronology level. The manuscript is well written and uses for the most part sound methodology. However, at times I think the authors introduce more complex analyses than are needed. The manuscript could be accepted after some minor revisions, but if the authors wish they could also consider some of the comments/thoughts provided below.
L103 Do you mean 20th century?
Figure 3 and 4 Please make sure that colorcoding is consistent between figures and species. PIJE and PIPO appears to be switched. This mixup could actually mask a stronger relationship between r2 and number of neg whiplash’s than is presently the represented.
L276 Not for positive events but perhaps for negative events..
L314 This is what you want to see but seems very rare in your dataset. Would it be possible to tease out more using PCA?
L328 This could be related to that you have treated the P12 and tree-ring data differently. I.e., applied detrending to the tree-ring data..
In general I think you have used a very elaborate reconstruction approach, that is, if I understood it correctly. Each site chronology is used as potential predictor with 2 negative and 2 positive lags, multiplied with the number of species chronologies. And yet only a handful are used to reconstruct water year, and surprisingly the chronology versions t0 is not in crushing majority.
Why not using a simple mean of each species and then regress P12 against each species mean and overall mean? Is this average not highly correlated with the water year? Would it be possible to tease out more using PCA? Is the first principal component carrying the water year signal?
How was collinearity treated in the regression models?
It would be helpful to plot the species means against the species reconstructions and the P12, at least in the SI.
Some of the figures seem repetitive (same results shown a bit differently) Fig 3, Tab 3, Fig 5.
The whiplash detection method also seems quite elaborated. Why not just use first differences and take the largest changes from year t-1 to year t? This way you will have treated both P12 and tree-ring data the same way. Right now it is difficult to determine if the detrending approach could obscure the detection of the “correct” whiplashes early in the tree-ring data around 1920..
An experiment for significance could be used to test if the methods of the authors is sensitive to overfitting. Generate random data that match the length of the original chronologies and lag these -2, -1, 0, 1, 2, years and then repeat the full method you applied. And do this X times to get distributions. If your tree-ring results are different from the 9X percentile of the randomized data, your models are likely not overfitted. I mention this because you seem to have a large variation in the chronology correlation with P12 and perhaps water year is not the leading limiting factor.
Would be good to have monthly climate calibrations for all chronologies to find out if the water year is actually a good target for the chronologies.
Author Response
Please see the attachment (uploaded pdf)
Author Response File: Author Response.pdf
