I have already commented on the first and second versions of this ms. In those last revisions, I was unsure, whether the results supported the conclusions. I am happy to say that this has now been amended. Since the last (second) revision the manuscript has gained clarity and I was able to follow the methods, except regarding the deseasonalization, and I agree with the fundamental parts of the discussion and conclusion. In my eyes, the ms still needs some work regarding clarity and rigor in the presentation, but it is now more editorial.
I recommend minor revisions.
Major comments
- Generally, the readability of the ms has strongly improved in this revision and it is much easier to follow the methods and the narrative in general. There are however still a couple of instances, where it is almost impossible to follow, for example, Line 701-727. I have indicated them below in the detailed comments.
- Some citations do not properly support the claim made. This needs to be improved. See details below.
- I have commented previously on the lack of explanation regarding the descriptive statistics and this has been improved. However, in most occasions where seasonal, e.g. twelve months, temporal coefficient of variation (temporal CV) is shown, the surrounding information is still incomplete. It is unclear whether the number refers to CV within this calendar month averaged across years or CV between values for this month between different years. Please add this information.
- I am missing a short statement already in the methods section of what information those CV carry. Are we expecting (or not) differences in CVs for ET between phenological seasons? Are high CVs an indication of uncertainty or are they expected in some seasons? This information follows very late in the discussion of the manuscript and the reader is a bit lost up to this point.
- The time series are „deseasonalized“ in a way that is not properly explained and unclear when following the references. It is not simply the calculation of anomalies from the seasonal average. But without understanding this, I cannot judge the results obtained based on the deseasonalized data, e.g. in sections 3.3.1 and 3.3.2. This can hopefully be cured easily, by shortly repeating in the methods section the basic steps applied.
- The ms, specifically the figures, mix two definitions of seasons, (a) based on climate-based seasons (e.g. Fig 15), (b) based on phenological seasons (e.g. Fig 11). Sometimes the same property is shown for both types of seasons (Fig 12, 13). In either case, there is no explanation for why one or the other option is chosen. It would streamline the paper if only one option was used throughout. Based on the title of the manuscript and discussion, I propose to only use phenological seasons.
Detailed comments
Abstract
Line 20: „flux observations“ better „surface exchange observations“
Line 37: „warm dry season“: It is a bit confusing from the reader’s perspective that the seasons are not discussed in chronological order. Therefore, would be good to guide the reader, e.g. "the preceding warm and dry season“
Lines 43-45: Here I expect a more general conclusion that directly relates to the title of the ms and possibly expands over the Miombo case. Is it appropriate to say that satellite products do not well represent ET over this data-sparse region with a phenology and seasonality that differ substantially from the typical case within the data-rich ground-truthing locations, this may also be the case for other locations with low data coverage?
Introduction
Line 59: „Intercepted radiation directly affects canopy conductance“
I disagree with this statement as written.
When stomata are closed, the radiation can be intercepted while canopy conductance remains low. I believe you mean that increased LAI enhances the transpiring surface and therefore canopy conductance. If yes, please reformulate and if no, please give a reference to support this statement. Also in the next sentence, it would be more correct to argue via the change in leaf area affecting the transpiring surface.
Line 63: „by moisture availability in both the vegetation ..“
In all environments I am aware of, the vegetation water content is not the main driver for transpiration. Are you sure that is the case here or is this maybe a typo? If indeed water storage within the vegetation is considered a substantial contribution to transpiration, this should also be accounted for in the discussion section.
Line 68-75: I propose moving the first sentence (L 68-70) to the end of this part (now Line 75), thus starting with the definition of evaporation and then addressing the evaporation in woodlands
Line 78: Please correct the reference to Roberts
Line 67-91: This paragraph is too long, pls split it up. For example, move lines 81 - 91 to a new paragraph
Line 85-88: The meaning of this sentence is unclear to me.
Line 82: „diverse“ unless you are referring to species diversity, I propose erasing this, as it is confusing. If you are indeed referring to species diversity, better write „species diversity“.
Line 86-87: „plant water storage“ - Why all of a sudden plant water storage comes up, while before only evaporation was in the focus? Is it important, can it be erased? If it is important, it needs to be introduced better. But since the remainder of the ms, including the discussion does not touch on plant water storage, I propose erasing it here.
Line 89: „ecohydrology“: Can you be a bit more specific? While evaporation is clearly defined above, the same is not true for ecohydrology. Therefore, it is unclear what is meant by this sentence. Maybe replace “ecohydrology” with “evaporation”?
Line 93-94: „which accounts for around 10% of the continent's land area“ - Great! Thanks for adding this.
Line 101-102: Erase „control exerted by the woodlands over“. Confusing wordiness, as vegetation reflects on phenology.
Line 123: Replace „The deep rooting characteristic“ with „Deep rooting, which is a characteristic“
Line 125-126: Erase „canopy serves as an evaporative surface that, in conjunction with other environmental factors, potentially“. Confusing wordiness, no information is lost when erasing this.
Line 124: Neither Fan et al. (2017) nor Kleidon and Heimann (1998) have worked on the Miombo, but the way they are cited makes it appear so. Please rephrase or remove the citations.
Line 129-130: „It appears that in the miombo woodlands, soil moisture increases with depth“ - This comment on soil moisture comes a bit out of the blue. Maybe move it up to line 124. the hydrology is introduced? Also, is this comment based on measurements?
Line 134-136: References are missing to the studies mentioned here. Especially, the comment on the carbon sink is surprising, since studies on an ecosystem acting as a carbon sink would be performed with eddy covariance data. And the latter would always also deliver information on evapotranspiration. However, it is mentioned earlier that almost no information on ET is available in the Miombo. This seems therefore a contradiction. Maybe specify the comment and also add some references to give examples of the existing studies mentioned.
Line 140: „on publicly available literature“ - are there further studies that are not publicly available? If yes, better state so explicitly.
Line 167-168: „these satellite-based evaporation estimates have primarily been developed for agricultural crops (i.e., Biggs et al., 2015; Zhang et al., 2016)“
I am not convinced by the references. Zhang et al (2016) also do not seem to agree. The terms „crop“ and „agriculture“ do not appear in the article, and the schematic Fig 3 in the paper shows tree canopies. Please either add other references or rephrase.
Line 170: I could not find this statement confirmed in Wang-Erlandson et al. (2016). Please double-check (also the other references, I do not have access to Snyder and Spano, 2013, but they seem more to be targeted at phenology, not remote sensing of ET in woodlands).
Line 183-186: This statement needs to be justified. The reference to one paper that makes recommendations for Europe („Bogawski and Bednorz (2014)“ ) does not serve the purpose.
Materials and Methods
Line 260-264: Split up this sentence, it is difficult to decipher.
Line 262: „contradicting “ is not the right verb here. Do you mean „to test“?
Line 302-306: Erase or add information? It seems trivial that e.g. leaf area index changes with phenology.
Line 314: „Identified five phenological seasons“
Here you are referring to season categories based on climate, not phenology Please omit the word “phenophase” here to avoid confusion. By definition this refers to the vegetation condition, which is not part of the classification explained in this paragraph. Also, it would be good to move the climate-based description before the phenophase description, because it explains the general frame in which the vegetation phenology is set.
Line 329: „ the comparison .. was conducted“ compared with what? please add.
Line 438: „serves as a practical solution“ - to what problem?
Line 475-490: It would be good to see the annual averages of the stations used to calculate the water balance separately because this would put the differences in the satellite products into perspective.
Line 536-541:
Please briefly explain the procedure in some sentences.
The first two references do not contain the words “adjusted seasonal factor” or “centered moving average”. I am not sure whether those terms are explained in the third reference since I have no access to it. In either case, the description is not sufficiently self-explanatory, to understand generally what was done and why it was required. Why not just calculate seasonal anomalies? Thus, for completeness, please add an explanation of the de-seasonalizing procedure. It should be possible to do so in very few words and would improve completeness.
Line 546 Erase „similar or“
Line 551 /Eq 2:
The coefficient of variation in the general form as stated here is very well known, and the equation can be erased. The problem with understanding, e.g. further down in Fig 8d is the lack of information over which sample (across years, within the same month) the mean and variation were computed. Please indicate this specifically on all occasions where the CV is used.
If you want to keep the equation, please move further up to follow directly Line 543, where CV is introduced. Its location in Line 551 is confusing here, as it does not relate to the Kendall correlation mentioned just above in Line 550.
Line 555: Unclear what „This“ refers to? The differences in CV?
Line 564 / Eq 3: Move up to follow „calculated“ in Line 564
Results and discussion
Line 607: „The study“ - As the opening of the section “This study” reads odd, when it refers, like in this case, to a different paper. I think you want to rephrase it to “A previous study”
Line 611: „field layer“ - I am not familiar with the term “field layer” do you mean “short structured vegetation” or „understorey“? I propose renaming or moving the explanation given in line 614 up here.
Line 616: „dry out“ - I believe you mean “enter a dormant phase” ? Or “loos leafs”?
Line 636: „trends“ - I believe you mean „variation“
Line 639: „highest mean LAI and mean NDVI“ - Not entirely. The highest NDVI in wet Miombo was observed in May in “senescence / green down”. Please reformulate to acknowledge this.
Figure 7: I am confused about the top photo with a date stamp of May 1: Above in Fig 2 and the Figure caption below you say maturity/peak is from January to March. But here the top photo with the time stamp May 1 is said to be peak too, but is later than stated above. Can you reconcile this presentation in the paper, either by adding a picture of March here or at least adding an explanation in the figure caption? Also, are you referring to the maturity/peak of the tree canopy only? If yes, would be good to add this information.
Figure 8: The legend in 8c is difficult to decipher. It needs to be larger.
Line 696: „Temporal coefficients of variation“ - Across years? A within the same month?
Line 7001-703: Please check the grammar of this sentence, something is wrong here, and it cannot be deciphered. Do you maybe mean to start with: “The time of peak LAI and peak NDVI” instead of “This Period of peak LAI”? Maybe use „corresponds to the NDVI peak“ instead of „align“
Table 2: Please indicate which metric for the correlation is given specifically. I believe it is correlation coefficients, but please specify. Also, it would be good to see the corresponding coefficients for the other seasons too, even if the correlation is low.
Line 718- 720: „temporal coefficients of variation“ and „ and Table 2“ - This sentence seems to relate to the values of Table 2. But in Table 2 apparently (see comment above), the correlation coefficients are shown, not the coefficients of variation. Please correct this.
Line 724: „temporal coefficients“ - Here as a reader I am lost. Do you now really refer to coefficients of variation? Please double-check and write it out completely to avoid confusion.
Line 779: „Fig A4“ - Figure A4 mentioned before A3?
Line 796: erase „non stationary“
Figure 10d-e: Using season adjustment when comparing the products and the canopy properties is very helpful. Please improve the description, however. As mentioned above, I do not understand the procedure for deseasonalization. I would expect some seasonal anomalies, but those would vary around zero, whereas here the values seem to indicate some average. It is hard to interpret. Can you please better explain what is the expected value and how the peaks and valleys in the time series can be interpreted?
Figure 11: Adjust the color scale to reach from 0 to 1 for easier readability.
Line 821: „anomalies“ - Which “anomalies”? The specification is also essential to understanding the difference between Figures 12 and 13
Section 3.3.2
Based on the information provided, I believe that only Figure 13 is required. It shows that there agreement between the different products shifts depending on the phenological stages. Since climate-based seasons also somehow reflect phenological stages, Figure 12 is not required.
Unfortunately, there is not enough information on what is meant by “anomalies” and by “deseasonalizing” so I cannot appreciate the the additional value of Fig 11. However, Fig 11 is not much discussed, and therefore maybe can also be omitted or moved to the supplement.
Line 882: „temporal coefficients of variation“ - Within years or across years?
Line 883 „means“ - spatial means?
Line 892-894: Erase sentence. Arguments before and after refer to how well products reflect ET during the dry season. This sentence is on a different topic and is distracting. I propose erasing or moving elsewhere, where absolute values of ET products are discussed.
Liner 896-899: The given information is not sufficient to understand this sentence. Do you mean “unrealistically high temporal variation in the estimates of evaporation”
Line 915: „the time series with seasonality revealed significant differences“ - I believe you mean „the original time series revealed differences ..“
Line 953-954: This is interesting. But how does the increased CV support this? Can you add a sentence?
Line 956: replace „differences in temporal trend“ with „differences in temporal dynamics“ as temporal trends were not analyzed.
Line 960: correct typo on „r > 0.5“ instead of „r > 5“
Line 981 - 988: „The green .. (Figs. 5,7 & 9)“ Erase. This has already been discussed above and does not add information here.
Line 991-993: „This implies .. dominates.“ - I am not convinced that this should be the case. The ET does not need to be affected by which layer is evaporating. Is there literature to show that ET is not properly estimated by satellite products during times when the understorey is more active than the overstorey? If yes, add references here.
Line 996 - 999: „This suggests .. dry season“ - Could you add a sentence, about why this variation indicates that the estimates are uncertain? ET could vary depending on weather conditions and differences in weather between years.
Line 1048: „estimates exhibited the most variation across phenophases“ If I understand correctly, you mean “is most distinct from the others”?
Figure 17: I find Fig A10 is easier to interpret than this one, but seems to contain similar information. I propose switching those two.
Line 1065: „Number of times“ - Difficult to interpret with absolute numbers, better give as a proportion of the total, e.g. 5%
Line 1078: After the end of the sentence, it would be good to add „Therefore, high variation is expected.“ It is really helpful for the reader to put the statement in context.
Line 1078 - 1081: „The temporal .. estimate“ - this is a repetition and can be erased.
Line 1112: „exhibit insignificant correlation“ better „showed no significant correlation“
Line 1138 / 1140 „over-year-storage“ better „interannual storage“
Line 1155: At „At the basin scale“ start a new paragraph and merge with the following paragraph.
Figures
The Figure captions are sometimes very sparse and the sub-panels are not referred to consistently. Please make sure all sub-panels are named (and use the same formatting, e.g. „(a)“), and also explain what is shown in the figure caption.
For example,
in Fig 4, 11, 17 the sub-panels are named by lowercase letters, e.g. (a) - (d) (which is good), but no explanation in the figure caption is missing. Please add this information.
The caption of Figure 5 mentions all the sup-panels (which is good), but they are not named in the Figure itself (which would be helpful). Please add names (a) - (e) both to the plots and the figure caption.
Figure 7 the dates of the different fotos are important to understand the text, but they have very small fonts at the bottom of each foto. Please use sup-panel letters and indicate the dates in the figure caption.
Figure 2: Please specify in the caption where the photos were taken.
Supplement
Figure A2: Something is off with the color scale. The correlation coefficients in A2 go from -1 to 1, with negative correlation accounted for. But the color scale only covers 0 to 1 with all values <= 0 falling into green. Would be good to also have differentiation for the negative correlations, some seem to be rather high, e.g. net radiation and soil moisture in the dormant season.
Table A1: caption: Please correct the typo in the second line to „2014, 2015 and 2016“.
Also, the line is missing below CFSR lowest part. |
