Effect of Provenance and Environmental Factors on Tree Growth and Tree Water Status of Norway Spruce

Round 1

Reviewer 1 Report

Title: Have You considered more climatic factors than environmental factors on tree growth?

95: Overall in this study has been focused on climatic variables and their effect on growth. However, the tree growth is also highly dependent on growth conditions (tree age, vitality, competition, past management activities etc). This should be also considered at least in introduction and in discussion.

111-112: Motivation and reasons why machine learning methods were used in analyses?

158-163: Measurement information of the forest stand, where the trees are located would be also needed to describe trees growing conditions. Hopefully measured information is available. Also more detailed information how trees were selected, and was there variation in their crown structure (vitality, growing position: dominant trees, suppressed trees) would be helpful for the reader.

147: CV => CW

216-219: Why the bands were located into 2.5 m height? Why the bark was removed, and could that have an effect on growth, spruce does not have very thick bark.

294: At what level the data was randomly split into training and validation subsets?

552: Also some discussion of the limitations of this study would be needed. Only one site, two origins, 10 trees measured for three-year period. How well the results can be generalized?

726-731: How about more replicates of origins and at different stage of stands? Will trees react differently at different development stage?

Author Response

Our responses to the reviewers comments (reviewer 1) are listed in the attached word file.

Author Response File: Author Response.docx

Reviewer 2 Report

The MS is a broad review account of available methods machine learning analytical methods (RF, ANN, etc.) with several phrases that need further discussion and which the authors take as widespread true. All these points are marked with a comment in the MS’s pdf file.

1.      (L309-462) The authors describe the figures (Fig. 3 – Fig. 7). I suggest the figures should be shown without text on them and all remarks should be written in the legend.

 

2.       What is meant by DOY in Fig. 8? Moreover, enlarge all figures since they are practically unreadable.

3.       Table 4 does not include probabilities that directly show the relevance of the models to the collected data.

4.       The selection of the Mini 32, EMS-Brno is excellent, but the authors must explain on what grounds their selection is based. Moreover, the selection of variables and the 4 methods (RF, GBM, SCM, and NN is never explained.

5.       In the Discussion section the authors repeat the description of the results and never explain their physical/biological meaning in the physiology of P. abies.

6.       I would ask the authors to put their ‘R’ command files in an Appendix or in Supplementary Material to facilitate the general reader to reproduce their analysis.

7.       Most plots are referred to as “variable importance plots” and partial dependence plots”.  This is not bad unless it is followed by the ecological meaning of this.

8.       In general, the authors follow the strategy of presenting in plots the partial dependence, the variable importance, and the statistics confirming the performance of machine learning models. The paper is full of plots related to 2 provenances (CW_PV or CV_PV, C_PV) × 11 variables (RF, DOY, ATavg, …) × 2 dependent variables (ΔW, MDS) × 4 ML models (RF, GBM, SVM, NN). This provides opportunities to make a lot of plots, tests, performance statistics and so on. To all these an ecological meaning is lacking for the reader, and he is directed only to the citations.

9.       In the “Conclusions” section it is stated (line 707) that “different growth strategies and mechanisms of two … provenances…”. This conclusion does not stem from the text.

 

I suggest rewriting, shrinking, and resubmission the MS.

 

Comments for author File: Comments.pdf

Author Response

Our responses to the reviewers comments (reviewer 1) are listed in the attached Word file.

Author Response File: Author Response.docx