the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 01 Jul 2025
Forest Diversity and Environmental Factors Shape Contrasting Soil-Litter BVOC and Methane Fluxes in Three Central Amazonian Ecosystems
Abstract. Biogenic volatile organic compounds (BVOCs) play a crucial role in biosphere-atmosphere interactions and the global carbon cycle. While vegetation is recognized as the primary source of BVOC fluxes in forest ecosystems, recent studies suggest that the carbon-rich soil-litter compartment contributes significantly. However, these fluxes, their underlying drivers, and their variability across forest types remain poorly understood, with measurements still scarce—particularly in the Amazon rainforest, the world’s largest source of BVOCs. In this study, we investigated soil-litter BVOC and methane fluxes and their potential drivers—including nutrient content, microbial biomass, soil temperature and moisture—across three forest types in central Amazonia: white sand forest (WS), upland forest (UP), and ancient river terrace forest (AR). Our results showed distinct flux patterns among forest types. WS exhibited both high emissions and consumption of gases, notably high acetaldehyde and methane emissions, and strong isoprene and monoterpene uptake. UP showed lower overall fluxes, with moderate emission and consumption of DMS, isoprene, and acetaldehyde. AR presented no significant fluxes. Linear models identified soil moisture and temperature as the primary drivers of fluxes in WS, while microbial biomass was the main driver in UP. Our measurements suggest that, despite covering a relatively small area in the Amazon basin, WS can be a significant ecosystem for BVOC and methane fluxes, regulated by soil moisture and temperature. Our findings underscore the need to account for forest-type-specific fluxes when modeling BVOC and methane emissions in the Amazon, particularly under changing climate conditions.
- Preprint
(1996 KB) - Metadata XML
-
Supplement
(1950 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-2895', Anonymous Referee #1, 01 Aug 2025
-
RC2: 'Comment on egusphere-2025-2895', Anonymous Referee #2, 15 Aug 2025
The manuscript “Forest Diversity and Environmental Factors Shape Contrasting Soil-Litter BVOC and Methane Fluxes in Three Central Amazonian Ecosystems” presents an extensive data set of surface gas (BVOC, CH4, and CO2) measurements across three forest types in the Amazon rainforest alongside a wide range of environmental, chemical, and microbial measurements. The data are valuable, rare, and were certainly collected with a tremendous amount of effort. The manuscript needs additional methodological clarification and organization/streamlining to clarify and highlight the findings. More information about the amount and type of litter contributing to the fluxes is crucial. This work will provide important new information to the scientific community.
Major comments
Nature of the soil and litter fluxes:
It’s unclear in the abstract (e.g., L27) and introduction whether this paper will cover flux of BVOCs between soil and litter (in which case we need to know the direction so we know what uptake or emission means in later parts of abstract), or their respective fluxes with the atmosphere. From the methods, it seems like these measurements are ‘forest floor’ and include both soil and litter inside the collar and thus represent the net soil/litter-atmosphere exchange. This should be clarified early on.
Soil vs litter contributions:
I do not find mention of the amount (mass or surface area) or plant species of the litter that was mentioned in the collars. Given the potential for litter emissions of VOCs (and maybe even uptake) to overwhelm soil fluxes, it is really important to indicate how much litter there was and how much it varied across measurements. I would suggest adding this as a table early on in your results. If these data are not available, this should be clearly stated and some indication of how consistent this is across and within transects should be presented.
Organization and streamlining:
The introduction is missing a motivation for the study. Rather than starting with BVOC sources and known drivers, it would be more compelling to motivate why study this at all. I suggest adding a big-picture section first which could include your L78-84. Then talk about specific sources/processes. Make sure motivation for studying different soil/forest types is clear, and goes beyond just that the Amazon has different soil/forest types. Please work on the flow of the motivating sections.
The results section could be more intentional about highlighting data that are most important in explaining the observed patterns. Other data could be moved to the supplement. For example, it would help to have the authors summarize and integrate the main findings in terms of potential drivers of gas fluxes. Listing the results in Tables 4-6 for each site is a bit overwhelming for the reader. These could be moved to the supplement or somehow combined and summarized. This is alluded to in L467-470, but I think the authors could also be more selective with the variables presented in figures and tables in the main text if they are not central to the findings and instead have them in the supplement. The role of the transects is not clear until section 3.3 where they are shown in terms of spatial variability. It would make more sense to me to present these results early on, alongside the site averages. I assume that this variability is considered in the comparison with the environmental, chemical, microbial data anyway. This would help reveal that there is a strong transect effect at WS likely driven by soil moisture earlier rather than later in the results.
The discussion should be organized around major topics (not order of results presented or analyses) and streamlined significantly. For example, rather than summarizing results by analysis in the discussion (e.g., separate section on PCA in L550 where you are repeating methods and results) I would summarize by major findings and discuss across your results. Try to remove repetition in the discussion with respect to your moisture and temperature results. Also, you relate isoprene emissions to microbes in two sections, and you could organize to discuss that in one location only.
Flux results:
Fluxes and their averages should be reported with associated uncertainty. It would be helpful to update the flux figures (3 & 8) in a way that makes it possible to see the fluxes of Up and AR sites in cases where the axes are overwhelmed by WS.
Minor comments
L24-26: This sentence reads awkwardly; suggest revising.
L27: It’s unclear whether this paper will cover flux of BVOCs between soil and litter (in which case we need to know the direction so we know what uptake or emission means in later parts of abstract), or their respective fluxes with the atmosphere. Please indicate.
L30-33: Unclear whether the summarized results are referring to both soil and litter fluxes.
L33: Of the variables tested, the models suggest these were strongest drivers.
L44: Are they a compartment together or each a compartment?
L46: What does essential mean in this context, please be more specific.
L52: Microbes also consume VOCs as part of their metabolism (not just release).
L53: CH4 is also consumed by microbes in soil. Clarify in the introduction how you are considering CH4 alongside BVOCs–is it or is it not a BVOC, and if it is categorized as GHG here, why is that important to distinguish? Why are you also measuring CH4 and CO2 here?
L70: This reference does not demonstrate this for BVOCs. Please make sure the sentence makes it clear what you are inferring from this paper.
L73-74: Are there additional references that can support the influence of vegetation? I imagine there may be others. It’s not clear how vegetation cover is different from plant species referenced in next sentence. Please clarify.
L84: It would be more helpful to cite the specific studies than this review.
L94: I don’t know that this is a helpful/informative way to start the section: “With a unique set of measurements”. It would be more helpful to motivate why the suite of measurements was particularly important for answering the questions.
L102: The emission/consumption rates of BVOCs, CO2, and CH4 (rather than gases). Check extra and? Here. List the drivers in (ii) that you tested (out of x, y, and z variables, what are the main drivers).
L141: Please make sure soil chamber measurements are described – we don’t know yet what these refer to. What are they and what do they measure? I don’t think it’s been clarified yet whether this is an in situ study or not, and this should be done before giving details on blanks etc.
L146: Using a probe?
L147: What depth of surface soil was collected?
L151: Meaning they were pooled and homogenized?
L152: We don’t know what these bag samples refer to. Please give an overview earlier.
L 155: I would suggest capitalizing ‘Transect’ and the names of the sites ‘White Sand Forest’ etc. as proper nouns. Also ‘Section X’
L163: Describe the collar materials and size. How deep were collars placed in soil? What does it mean to ‘seal with the surrounding soil’ in L170, please also describe here. Was the effective volume 21L including the collars? If not, please give range of volumes with collars.
Fig. 2: The teflon line is illustrated with an arrow out. There is an apparently empty fitting on the top of the lid. Which of these show the air inlet to the chamber, and could you have arrows indicate the direction of the gas flow? The use of arrows to show clamps, soil collar, O-ring, etc seem counterintuitive to have the arrowhead point to the label instead of the named item. Are the litter and soil positioned above the clamps (entire soil collar buried) or could you be more specific about their position within the soil collar and the collar’s vertical dimensions in general and wrt the soil?
How was the blank chamber bottom ‘completely sealed’? Does it matter if the blank and soil chambers had different internal volumes because the soil collar was partially in the soil?wu
Consider merging sections 2.3 and 2.4.
L176: The theory/rationale behind the sampling approach should be presented. Clarify whether at this volume and flow rate the approach is considered closed/static or flow-through? The flux chamber would not have been at steady state after 20 mins (10 min*0.5 LPM = 5 L removed out of 21 L, so headspace had not turned over even once). So was the goal to sample gases that had accumulated after 20 mins? Please specify in more detail.
Could you have detected gas/BVOC uptake by this method? Please also specify. You do report negative flux values – do you trust these as indicating uptake? Please state.
L186: This is the inlet line to the chamber? Please clarify.
Section 2.5. Could you indicate whether this analyzer can be used for high-resolution analysis of mass spectra or whether the results are all analyzed at unit mass resolution for this study? Do the specific masses need to be chosen in advance, and they are chosen at unit mass? If so, how were they chosen? A little background info on this will help.
L226-241: Much of this section may be better suited to the discussion. I understand that it is providing rationale for some of the methods, but it also includes the discussion of your results. Reconsider placement.
L252: I would suggest performing a sensitivity analysis to prove this. You could take the concentrations from the blank and the dilution information from your flow rate and chamber/aboveground collar dimensions and estimate it.
L308: Please clarify how you propagated uncertainty to arrive at 3-5 significant figures in your flux results. Add a measure of variability / confidence to the reported fluxes (stdev, ci, etc) throughout the manuscript.
Fig. 3: The yellow rectangle is not necessary.’ It would be helpful to indicate the ‘zero’ line in flux plots. The fluxes for many VOCs for AR and Up are indistinguishable from zero in this plot. Could you indicate if they are significantly different from zero, and if so, find way to also visualize them? Cases of uptake and emission from these sites are mentioned in the results (e.g,. L324, L309), but without quantification and we can’t see them on the figures.
L329-332: This is a discussion point.
Fig. 4 and Fig.5: I would suggest using names (not shorthand) for the figure titles (Aluminum instead of al_soil, pH instead of ph_soil). If the caption says soil or litter, you don’t need to add to each title.
L362: This first sentence should be qualified by saying ‘only for P and for N but only in litter. You are not showing significant differences between forest types in most cases.
Fig. 7: It would be helpful to construct this PCA for all sites (as shown) but also for each site individually. This could help you explore potentially important variables associated with forest type or variables within a forest type, and whether they align in the same way with the various gases. For this figure, were gas fluxes (or other variables) transformed (e.g., center log transformed) in some way to account for any non-normality?
L369: This acronym was already defined, and is defined at least 3 times in the paper.
L472-491: It is not clear how these findings (e.g., soil iron) relate to your questions in the introduction related to gases. It is not brought up again. I would focus the discussion on topics pertaining to the gases.
L498: These references pertain to other sites, right? Please clarify they aren’t references to your sites. Could you discuss the potential drivers in terms of your results first, and then later you could consider other factors you didn’t measure as other drivers that may explain additional variability?
L515: Do you mean soil uptake or soil and litter uptake in your study? Could the difference be due to isoprene emission from your litter? Please consider the absence/presence of litter when you compare to other studies. Pugliese et al., 2023 also reports deposition velocity so you can compare the concentration-independent uptake as well and determine whether there a difference persists when you correct for the atmosphere.
L546: I don’t know if attributing to ‘natural variations’ is particularly useful.
L569: Could you expand more on why this may be? Why would the response to soil moisture differ in different soil and forest types?
L582: This is not a strong way to conclude your discussion paragraph here. The temperature only varied by 2C across all your sites, which is very little. You could calculate the sensitivity of expected soil VOC emissions to that temperature change based on other papers, and I think it would be small. More than temperature, light penetration to the surface may have an effect. Could you discuss this and try to arrive at a more specific conclusion even if is just for 1-2 gases?
In section 4.3.2 please be very clear where the studies you site are for plant emissions of VOCs versus soil or litter studies. I would suggest streamlining this discussion and being very clear about how previous work informs the fluxes you observed from soil and litter.
L653: Making measurements on equidistant points does not influence the inherent spatial heterogeneity of a system. Please clarify in methods how you selected homogeneous areas–how and why was this done?
What drives the differences between forest types? How could forest-type-specific fluxes be inferred, or is the recommendation to go inventory all forest types?
Section 4.5: This section should be streamlined to focus on the results more than making a separate broader point that would be more well suited for a commentary or opinion piece. The point is already made succinctly in Section 5.
Supplement
Is there a discrepancy in the sampling rate, and thus total volume sampled, between section 2.4 and the supplemental methods section 3? Confirm that the determined concentrations accounted for the air volume sampled.
What is the purpose of Fig. S2? Please include uncertainty bars. How are these results included in the main manuscript?
What did the blank chamber measurements look like? Were concentrations fairly steady across all the samples?
4.1: I don’t understand the difference between Fig. S3a and S3b is even after reading the text a few times. Would you please make this more obvious?
Fig. S6, S7: What is the p-value on this correlation? Is it significant? These differences in temperature are small.
Citation: https://doi.org/10.5194/egusphere-2025-2895-RC2 -
RC3: 'Comment on egusphere-2025-2895', Anonymous Referee #3, 19 Aug 2025
The manuscript by D. Pinheiro-Oliveira does an attempt to quantify CH4 and BVOC emissions from different forest ecosystems in the Amazon. Doing so gas samples were collected at one point in time (December 2021), along duplicated transects in three different forest types. Along with BVOC and CH4 data, data of potential drivers were collected.
In principle this are interesting data. Despite no major methodological flaws are presented in the paper/with the data set, I have several concerns. These concerns are listed below, both in general and specifically.
First, the data set has a very limited temporal coverage, i.e. one time point that has been repeated the next day for a spatially duplicated transect in one specific period, i.e. December 2021. As a result, one forest type has 12 samples from a very specific period. This also is the case for dynamic parameters, such as soil moisture and temperature, which are used as drivers to explain the observed BVOC and CH4 fluxes. As a result, this rather long paper is loaded with sections in the discussion that are very speculative (see examples below) that are not fully supported by data. A next step is than a danger of overstating some conclusions (see examples below) in the discussion. Indirectly the authors mention this limitation in the description of the statistical methods where they reduce the number of explanatory variables due to low amount of data per forest type, i.e. 12. In addition to this, it is also unclear using the PCA in figure 7, how predictor variables were finally selected/retained.
Second, the method used to calculate BVOC/CH4 fluxes, i.e. collecting a large volume of gas and using a blank chamber is unusual, but not wrong per se. Likely this is needed for the large volumes that need to be collected to detect BVOCs with mass spectrometry? This could be better explained/justified in the text.
Third, another issue is the way nutrients were extracted, i.e. using nitro perchloric solution. The latter is HNO3+HCl. This suggests total nutrient contents were measured and not bio-available levels. There is also a poor discussion on why these nutrient levels are what they are (section 4.1).
Fourth, if living roots are also a potential source of BVOCS (line 508-511) why wasn’t a root exclusion experiment added as an additional treatment? This would have made the data set, that has a very limited temporal resolution, more interesting.
Fifth, I think more efforts could have been done resolve this issue with m/z 42. Was this not solvable?
Some specific comments:
Line 71: I wonder that you can really generalise that CH4 fluxes are higher from sandy soils? Or do you refer to CH4 uptake fluxes (i.e. CH4 oxidation)?
Line 78: How can you state the Amazon is the largest source of BVOCS, without existing data from the Congo basin? Mention also the role of BVOCs for O3 and NOx dynamics.
Line 225: no good fit for acetone, ethanol and formaldehyde,… what does this exactly mean?
Line 308-309 (as example): no need to use 3 decimals here.
Line 329-332: You cannot make a “translation” to “soil moisture available to plants” without soil physical data such as a Pf curve.
Line 462: higher than anticipated, how much higher?
Line 516: I think this is rather a diffusion into the soil, then an effective (microbial) or (physical) uptake.
Line695-696: how can a forest type with an aerial coverage of 5% offset all carbon losses from other forest types?
Examples of “speculation”: lines 482, 487, 507, 530, 550-559 (the entire soil moisture and temperature issue as drivers for CH4 and BVOC fluxes, cannot be covered with this limited temporal data set!), 575-577 (idem), 582-584 (idem), line 675 (due to the low temporal coverage the extreme conditions of especially the white sand soils could not be covered),...
Examples of “overstating”: lines 36, 468, 568-569, 692,…
Citation: https://doi.org/10.5194/egusphere-2025-2895-RC3
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 679 | 79 | 20 | 778 | 23 | 14 | 23 |
- HTML: 679
- PDF: 79
- XML: 20
- Total: 778
- Supplement: 23
- BibTeX: 14
- EndNote: 23
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
In this paper, the authors studied the influence of forest diversity, environmental soil factors, and microbial biomass on soil-litter BVOC emissions and GHG fluxes in Amazonian forests. Three different ecosystems were selected, and a broad range of drivers was analyzed. Two GHGs (CH₄ and CO₂) were measured, along with various BVOCs. This paper presents novel research in a field that requires further investigation. To date, limited research has focused on BVOCs from the soil compartment of forests. This is important, as the Amazon contains the largest tropical forest in the world, and global BVOC emissions are predominantly from natural sources. Despite the extensive number of variables measured and therefore, data obtained, the authors made a selection for their main text.
General remarks:
Overall, the authors should ensure consistency in the use of abbreviations; for example, "CH₄" is used in some instances, while "methane" is used in others. The same applies to the naming of forests, as well as to abbreviations like DMS and LMs.
Regarding the use of Tedlar bags, do you expect any sorption/losses of gases when using these bags? Was this checked for the compounds of interest, or do you have any literature that supports this?
N2O is also a very important GHG emitted/consumed by soils; however, it is not mentioned, not even in the introduction. I wonder why the authors chose not to measure it.
I am not sure if I overlooked this, but is there any explanation for the patterns of isoprene and monoterpenes along the transects?
Lastly, I am curious whether the authors were able to identify other BVOCs in their samples that could be relevant for future studies. Based on the sampling strategy and analytical techniques used, I suspect that more compounds were observed and they could be tentatively identified using an MS library. This is, in my opinion, of great added value for future studies that decide to expand (in advance) their current list of BVOCs of interest, especially given the limited number of studies on the subject.
Specific remarks: