Phenotypic Plasticity of Drought Tolerance Traits in a Widespread Eucalypt (Eucalyptus obliqua)

Round 1

Reviewer 1 Report

General Comments

I find the work to be a valuable contribution to Forests. The paper is nicely written and illustrated and I congratulate the authors for presenting work that monitors how a widespread eucalypt species deals with seasonal variation in environmental conditions. Additionally, I applaud the authors for being forthcoming that the throughfall setup has not yet managed to dry the soils. In my view, just the data and analysis alone of seasonal variation in water relations of the target species is already worthy of publication, as it provides an important basis for understanding how drought will affect physiological performance in a really severe drought. The description of the throughfall set is also beneficial as it can be a foundational for future studies using the infrastructure.

There are however a number of considerations that guides my suggestion of a minor revision for the paper. 

Firstly, the way the throughfall experiment has been presented in the paper makes me wonder if there was a better way to present this. In a large part, this is because of the very conspicuous omission of data detailing the effect (or lack thereof) of the infrastructure. While the authors claim that the throughfall setup was ineffective in causing the drying of the soil (line 13-15), I think that it is important to present the data for readers to see. I presume (hope) you used soil moisture probes to monitor the soil volumetric water content (VWC) of the soil over the 2.5-3 years while the throughfall infrastructure was in place. (e.g. See Limousin et al. 2008) You can then present this VWC in a set of boxplots and ANOVA analyses in the Supplementary Material.

The second major issue is how you have defined "drought". You have presented good data and informative figures showing precipitation, temperature and a calculated heat moisture index. However, from the literature related to plant water relations, it seems that drought is more typically related to soil water deficit and vapour pressure deficit (See Eamus et al. 2013; Grossoird et al. 2020). In any case (I might be wrong) but these are some of the factors that would have the most direct relevance for plant water relations measures such as water potentials. Again, having soil data is quite critical to establishing a soil "drought" status. Throughfall experiments can really only modify soil moisture (maybe at least in the first few years), so with no soil moisture data presented, there is no basis for saying whether it did or did not work. Also, given the way the meteorological data is presented, there are also no good reasons to believe that the naturally drier summers are "drought" conditions. I was wondering if it was possible to provide some of these data (e.g. soil volumetric water content, and if possible vapour pressure deficit or relative humidity) as this would bolster the premise of a "drought" and in any case bolster the lower water availability to plants. Alternatively (and perhaps preferably), the authors could forego the use of the term "drought" in this paper but frame the paper more as a study of plant physiological (water relations) responses between seasons. In lieu of "drought", perhaps it would be more accurate to use "dry season" or "summer". Note than later you mention a predawn leaf water potential value (more negative than -0.75MPa) where you consider to be droughted conditions, but this does not really match up with the P50 values you presented in Fig S1 which was almost -3.0MPa. Even the most negative midday leaf water potentials you presented are not more negative than 2.6MPa. Given the previous P50 value you found in your other work, I do not know if I would consider your plants being in a drought (See more comments on this later). By the way, since you did some extensive P50 curves with the species, since you have P50/P88 values for individuals sampled at this current sampling site at Wombat reserve, it may be useful data to highlight in this paper.

Have you also considered looking at the difference between the midday and predawn across time? One hypothesis would be that the throughfall treament plants would have a greater midday-predawn difference than the control plants. Just a thought. For an example of what I mean, see Figure 2a in Martin-StPaul et al. (2017)

In conclusion, I do not think that the issues I have raised are unfixable, and I would like to just reaffirm that I feel that the data and analyses presented in this paper will make a valuable contribution to the plant hydraulics literature. If you are able to fix these issues, I would be very keen to review a revised version of the manuscript.

Data analyses
You should state explicitly somewhere that for the purposes of some analyses you pooled the trees from all the three throughfall plots as trees with throughfall treatment and Control plots as Control treatment. This way you circumvent questions on whether the individual plots had an effect.
The way you have analysed the data throughout a big part of the paper (based on the results you presented in Figures 5, 6 and 7) suggests that you used all the trees (both in the throughfall and control). That is fine (since the throughfall had not yet started to show effects), but you should perhaps state it explicitly somewhere in the Methods or data analysis section that because the throughfall had not yet started to show effects, you used all the trees (pooled from both in the throughfall and control) for analysis of seasonal variation.
Note that I have mentioned somewhere else in the review that you could perhaps already present the fact that your throughfall setup did not sufficiently dry the soils in the Methods, and then use that as a rationale to state that you will just be analysing seasonal variation using all the plants.
The other option for the analyses you presented in Figures 5, 6 and 7 (which may be preferable depending on how you frame it) is to do analyses of only the subset of trees in the three Control plots. You lose replication but 9 trees is still a respectable dataset (especially since you repeated the measurements over 3 years on the same trees I presume). If you choose this approach, you can ignore any rogue effects you may find in the throughfall plots, and also, the data will be much more readily comparable to other datasets.
Regarding your use of ANOVAs, I wonder if it would be more accurate to have a nested design of season nested within year, or a repreated measures design. I am not sure if doing multiple one way ANOVAs is the most appropriate way to analyse your data. You may need to explore glms or mixed models where you can nest season within the year. The effects of the individual seasons can then be determined by a posthoc test. Ultimately, I do not think that is going to change the result, but you will likely be presenting a more appropriate data analsyis.
Also, you need to specify and reference the program you used to do your data analyses.

Figures and captions
Figure 1 could perhaps be stacked on top of Figure 3. Worth considering if you need to save space on a figure.
Figure 1 and 3 captions should probably include the study site.
In the captions for Figures 5-7, it would probably be good to mention the number of trees you used. Figure 5's caption could also specify the species you used.
In the captions for Figure 2 and 3 you cited the subfigures in complete parentheses, e.g. (a), (b) etc. However, in the captions for Figure 4-7 you used incomplete parentheses., e.g a), b). Please standardize across all figures and also in the graphs and the captions for all figures. Probably good to use complete parentheses.
In Figure 3
For figure 6a, it would probably look more consistent if you placed the "a)" inside the bounds of the graph, especially since b) is within the graph

Page 1
Line 11 multi-year droughts or just a natural reduction in water availability in summer? I think you will have to define droughts better in the Intro/Methods. Additionally, you do not actually show in your work that a lower water availability in one year affects the next (i.e. have a compunding effect), so maybe what you more acurately tested is seasonal variation over three years
Line 18-19, 23 - Why only predawn leaf water potentials? The midday leaf water potentials (wich you meassured) are generally (and as expected) more negative and so wouldn't that tell you something about the negative pressures the species can operate under. See other comments elswwhere.

Page 2
Line 52 - In addition to Martin-StPaul et al., 2013 and Tng et al., 2018, perhaps you should also cite and look at Jennings et al 2017. They did a 50% throughall exclusion experiment, probably using a very similar method to yours. Their work is just presented as an Absteact but it might be insightful.
Line 80-82 - I do not think your current study design allows you to answer examine the effects of "(ii) long-term reductions in water availability", but your first aim to look at the effects of season is very defensible.
You should perhaps also state your hypotheses/predictions with regards to the throughfall experiment such as "We hypothesize that the plants exposed to the throughfall experiment will exhibit decreased XXXXX, XXX, XXX." Even though the reader knows by now (after reading your Abstract) that your throughfall had not yet started showing effects, it still makes sense that you have an initial hypothesis about it, since you dedicate whole sections in the Results and Discussion to the topic.
Line 88 - Could you also specify the slope of the site? I imagine for that lateral flow to be of an issue the slope may matter. As will soil porosity. I doubt it anyways, unless there is a history of flooding in those areas. In any case I think you should describe the site better, e.g. soil type, aspect, height of dominant and subdominant trees, presence of a shrub layer (and it's properties, i.e. broadleaf or sclerophyll shrubs? or a mix.). Is there also a specific vegetation type name you could cite? I imagine that Victoria's veg classification system is pretty well developed.
Line 93 - Perphaps you could state explicitly that the site experiences dry summers and wet winters, or some version of the majority of the precipitation occuring in the winter months.

Page 3
Line 98 - Do you really need a abbreviation for Standardized Precipitation Evapotranspiration Index (SPEI). You use it only 2 tinmes in the text in Line 95 and Line 101.
Line 99 - The longitude you presented does not have any range.
Line 101 - It is not great to start a sentence with an abbreviation. See previous comment. I recommend spelling SPEI out in full.
Line 104-105 - I am not sure what you mean by "our study site at 1015 mm mean annual precipitation reflects the growing conditions for many
individuals of the species." Did you mean the near-optimal growing conditions?
Line 109 - Indicate how many records of Eucalyptus obliqua you used in total. Indicate also where you got these records from, i.e., Atlas of Living Australia.

Page 4
Line 123-124 - You have based your "drought" conditions on the heat moisture index, and although it is probably very justifiable to say that a low HMI means a period of very low moisture availability, you have not cited any supportive literature to say that low HMI = drought conditions. The low HMI that you have found in all the summer seasons between 2015-2019 coincides (naturally) with low precipitation and high temperature periods. These are natural cycles and may not reflect real drought in the sense of a drought that kills trees.

Page 5
Line 132 - Could you mention the size range of the trees you measured. I know you mentioned this in line 88-89 but this is not specific to the individuals you measured Could tree size potentially be used as a fixed or random effect in your models.
Line 133-135 - The description of the thoroughfall experiment could benefit from a figure. This figure could go into the Supplementary Material. I think it would help your readers see how the rainfall exclusion was achieved
Also, you mentioned rainfall exclusion gutters, but were there not panels to filter the rain down into the gutters? Or was it all just gutters between the boles of the trees. You also said that the gutters covering only 50% of the plot. Does that mean that the remaiaing 50% of the plot is totally uncovered and that rain still gets to the soil in those uncovered areas? Doesn't that mean that even within those 3 plots you chose for the "drought" plots, half of their area actually still gets rain? I think you need to describe better the thoughfall exclusion method and illustrate wth images (at least in the Supp. Material). I went to the drought net website and to look at their protocols, and what I found was that there are shelter methods (e.g. like those used at Manaus, Brazil and the Daintree) and also fixed troughs. From your description, it sounds like you are using the latter. Again, it all becomes clear if you provide a little figure (e.g. See Fig. 1a in Martin-StPaul et al. 2013). Maybe such a photo and location details of the Wombat site could be worked into your Figure 2. If this is the case, you could perhaps also cite some specific literature that have used the fixed trough system.
Line 136-137. You should perhaps also specify how many trees you used in total. I presume 3 trees x 6 plots = 18 trees? 9 from the throughfall and 9 from the Control? In your figure captions, it would be also useful to mention the number of trees you used.
Line 138-139 - Could you also specify the position (height, canopy position, exposed branches?) of the trees where you collected the shoots. Usually this is done in the outer canopy on sun exposed branches with healthy leaves.
Line 144 - healthy fully expanded leaves?
Line 157 - The dry weight was recorded after oven drying leaves for 48 h at 65 °C or until a constant weight was achieved.
Line 159 - One way ANOVAs were used to test.... (However, see my comments above on your data analysis)
Line 169 - Why do you test the effect of HMI on predawn leaf water potential only. It is not made clear anywhere why you chose to use only predawn. Doesn't it make sense to test it for midday leaf water potential as well? Likewise, shouldn't you also test the effect of midday leaf water potential on the other functional traits as well. Note that when you relate your results to Fig S1, the leaf waster potential that makes the most sense is the one that is the most negative.
Perhaps it is ok to use only one, since the predawn and midday should show similar trends, but maybe your Fig 7a showing the tight correlation between predawn and midday leaf water potential should be in the data analysis section to help justify this. You can then make a case there for why you chose using only predawn leaf water potential.

Page 7
Line 187 - I suppose when you say impact sites you mean the plots with the throughfall exclusion infrastructure? It is probably better to be more explicit about that in a Figure caption. Also for a and b, you should perhaps specify you mean leaf water potentials.

Page 9
Line 208-209 - Firstly, this statement you made here sounds like something that belongs to the discussion, and not so much the results. It makes more sense for you to discuss this result in the Discussion section, and also place it in the context of Figure S1 that you provided. You have done some of this in Lines 279-283, so why not discuss your predawn LWP homoestasis concept here. As a matter of interest, it would be prudent to note that the P50 figures presented by in your other work Pritzkow et al. suggests that you have not yet come arrived at a water deficit level that approximates the Midday LWP of their P50. This being the case, I do not know whether you can conclude that your plants are droughted. By the way, I have looked at your 2020 Tree Physio paper, and it looks like you have sampled trees from the Wombat population, so why not use that data for that population as well (maybe in Fig S1 you could use different colors for the dots to represent trees that are from the Wombat population, or maybe just use the Wombat population trees to plot that graph)
Line 213 - All figure captions should be understandable as they are standalone so I think you should specify that you mean leaf water potential

Page 11
Line 246-247 - If only 50% of the plot was covered by the troughs (or gutters), I am not sure whether trenching would be of any consequence. The rain would still get to the middle 5m x 5m of the plot where you sampled. Still, I wouldn't be too quick to say that your throughfall setup was ineffective. Perhaps it just hasn't managed to dry out the area enough for you to see significant changes. Everything will depend on what soil moisture data you can present on the throughfall setup

Page 13
Line 366 - italicize Eucalyptus grandis, do this also for scientific names in lines 383, 385, 409, 443, 465, 495, 501, 510

Page 14
Line 402 - you are either missing the volume or issue number
Line 413 - italicize Quercus ilex

Page 13
General comment: Check whether the journal really needs issue number
Line 452 - italicise the scientific names

Supplementary Material

"Supplementary" should be "Supplementary Material"
Caption for Table S1 - See line 4. "leaf are" should be "leaf area"

The caption for Table S2 should be given a space after the end of Table S1

Figure S1 caption - spell out Eucalyptus obliqua in full. Also check if there are copyright issues with re-publishing data that has already been published elsewhere. In any case, since the Supplementary Material is a stand alone document, if you were going to cite a paper, you need a reference section so readers know the full citation of Pritzkow et al, 2019. Also, you need to provide the number of measurements (n = ??) Does each dot represent a single population?

References to consider

Eamus, D., Boulain, N., Cleverly, J., & Breshears, D. D. (2013). Global change‐type drought‐induced tree mortality: vapor pressure deficit is more important than temperature per se in causing decline in tree health. Ecology and evolution, 3(8), 2711-2729.

Grossiord, C., Buckley, T. N., Cernusak, L. A., Novick, K. A., Poulter, B., Siegwolf, R. T., ... & McDowell, N. G. (2020). Plant responses to rising vapor pressure deficit. New Phytologist, 226(6), 1550-1566.

Jennings, K., McIntire, C., Coble, A., Vandeboncoeur, M., Rustad, L., Templer, P., & Absbjornsen, H. (2017). Tree species' responses to throughfall removal experiments superimposed on a natural drought event in two contrasting humid temperate forests in New Hampshire, USA. EGUGA, 19461.

Author Response

See attached word file for details

Author Response File: Author Response.docx

Reviewer 2 Report

A fairly interesting work on multi-year physiological and morphological responses of adult eucalypt individuals to water availability dynamics in the Wombat State Forest! Key parameters reflecting irregular occurrence of summer droughts were analysed here. Unfortunately, the experiment on the simulation of chronic drought (throughfall reduction experiment) was not successful (as I suppose –  particularly because of still high precipitation in the study area). In spite of this, there are a few issues, which should be considered:

1. I would add Latin name of the eucalypt species in the title.
2. Please, avoid repetition of formulations in following sentences (e.g. lines 11-12, 39-40).
3. I find the description of meteorological conditions in the Material and Methods, already presented in graphs, as not necessary.
4. Please, check width of the shaded areas in Fig. 3. Shouldn´t they indicate respective years?
5. Please, define, which leaf from what position in the tree crown was submitted to the functional trait measurements.
6. Since changes in SLA were associated not only with the appearance of drought but also with growth, I recommend to present important points of the tree phenology, as well.
7. Line 174 – correct to “throughfall reduction plots”.
8. Fig. 4 caption – to be clear, please, replace “impact” with “throughfall reduction”.
9. Line 217 – relations between HMI and water potential parameters - except for Ψ_PD - were not presented (please, add them or do change in the text).
10. Line 261 – “…but greater than observed for other eucalypts…” – is there any literature, explaining this difference?
11. Supplementary material: Table S1 – replace “impact” with “throughfall reduction”,
12. Supplementary material: Fig. S1 – is the reference Pritzkow et al. (2019) correct? In the list of references I found only Pritzkow et al. (2020).

Author Response

See attached word file for details

Author Response File: Author Response.docx