the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 30 Jul 2026
Canopy processing of atmospheric dust creates a short-term phosphorus pathway in a tropical forest
Abstract. Phosphorus (P) availability constrains productivity in many humid tropical forests, where highly weathered soils retain P in poorly available forms and external inputs are needed to offset long-term losses. Caribbean forests receive new P through trans-Atlantic Saharan dust and episodic African biomass-burning particles, however how this particulate P enters forest nutrient cycling remains poorly understood. Here, we examined the fate and uptake of particulate P in the Luquillo Experimental Forest, Puerto Rico, using a foliar dust-application experiment and soil incubations with mineral dust and wildfire ash. In soils, both materials increased P supply to ion-exchange membranes, but mineral dust-derived P was rapidly transferred into Fe- and Al-associated pools, whereas wildfire ash generated greater short-term P availability. On leaves, mineral dust exposure resulted in foliar uptake of dust-derived Fe and P, with P uptake quantified using Fe as a conservative anchor. Particles recovered from leaf surfaces contained more bicarbonate-extractable P than the applied dust, indicating that contact with leaves increased the lability of P remaining in deposited particles. Together, these findings identify two complementary canopy pathways: direct foliar uptake of dust-derived P and leaf-surface processing that increases the fraction potentially available for subsequent uptake, either from foliage or, following redistribution in throughfall, by roots. By revealing a pre-soil pathway for atmospheric P, our results highlight a process not explicitly represented in most terrestrial nutrient-cycle frameworks that may affect how models represent the biological availability of dust-derived P and its coupling to tropical forest carbon cycling.
- Preprint
(1291 KB) - Metadata XML
-
Supplement
(43 KB) - BibTeX
- EndNote
Status: open (until 28 Oct 2026)
- CC1: 'Comment on egusphere-2026-4320', Charles T. Driscoll, 03 Sep 2026 reply
-
RC1: 'Comment on egusphere-2026-4320', Anonymous Referee #1, 04 Sep 2026
reply
The manuscript (MS No.: egusphere-2026-4320) “Canopy processing of atmospheric dust creates a short-term phosphorus pathway in a tropical forest” by A. Gross et al. describes an interesting experimental study in which the authors (sort of) compare foliar uptake of phosphorus (P) from atmospheric particles with soil processing of P from atmospheric particles applied to soil. The authors are comparing apples with oranges but that notwithstanding, this is a challenging question to address. I like the paper and the authors should be commended for taking this research on. The authors provide a unique perspective on the important role of foliage in the processing and assimilation of atmospheric particles to obtain P. A challenge for the authors is to present a clear and coherent description of a complex experiment which contrasts foliar processing with soil processing of atmospheric particles using contrasting methods. Although I like the paper and think it should be ultimately published, I had much difficulty following the author’s text. I found myself going back and forth through the text trying to follow their experimental approach. I don’t think there is any way around this problem. The authors have undertaken a complex experiment and there are clear limitations, but the results are interesting and the paper could ultimately be published in EGU Sphere.
One of the big issues I have with the paper is understanding how the dust is “applied” to the leaves. The authors state that they “use a fine sieve to apply uniform coverage” But how is this done? The dust is a dry material. What keeps it from falling off the leaf after it is applied? Of the mass of dust applied to the leaves what fraction remains on the leaf? This experimental application should be clarified on p. 9, line 242.
Another concern that I have is that it seems that the mass of dust applied experimentally is quite large; approximately 70 g m-2 of leaf surface. The authors state that is approximately equivalent to 3 years of deposition. I understand that the authors would like to apply a large dose to observe a plant response, but could there be limitations associated with over dosing? If so these limitations should be discussed in the Discussion section of the paper.
I don’t like the fact that the authors made two separate applications of two different types of dust to the same leaves. I think this is a serious flaw in the experiment. How can the authors determine if one type of dust reacts substantially differently to the leaf than the other type. I understand that the authors had limited quantity of atmospheric dust, but this approach is a clear limitation to interpreting the experimental observations.
In various sections of the Discussion the authors address some of the limitations of the experiment. However, they don’t discuss the two limitations which I mention above and I consider these to be major limitations of the study. I think these limitations should be addressed in a separate “limitations” section of the Discussion which would include all the other limitations mentioned throughout the Discussion section.
I have a number of smaller points and comments which should be addressed before the manuscript is suitable for publication.
Line 133 I don’t understand “dry deposition is approximately 53 -73%”. Of what? Is this percent as P or total mass? If it is total mass, how is total mass expressed? Dry mass of wet plus dry deposition? The authors need to clarify what they are trying to say here.
Lines 177-180 How are the local dust and ash collected. Are these surface soil deposits? What depth? Some additional detail on the collection of these materials is needed. How soon after the fire was the ash collected?
Lines 261, 270, 379 and probably others elsewhere The authors cite Lokshin et al. 2026 and Lokshin et al. 2026a in the text. In the references they provide Lokshin et al. 2026a and 2026b. However this seems to be the same article cited twice. Something is incorrect and the authors should fix it.
Line 302 The authors should clarify here that the greenhouse experiment is Starr et al. 2023.
Line 393 It appears that the authors cite Figure 3a here is the text before they cite Figure 2. If I am correct the authors should probably switch the order of Figures 2 and 3.
Line 405 Should equation 3 be %ΔPest ?
Line 476 Why do the authors use ppm to express concentration units here when they use mg P g-1 elsewhere in the paper?
Line 484 The authors are not really normalizing the P fluxes to the soil surface area, are they? Aren’t the values normalized to the soil pot surface area (0.01 m2)? The soil presumably has a relatively high specific surface area and therefore a high surface area. This description should be clarified in the revised text.
Line 490 and Figure 4a I am completely confused by the normalization of the foliar P response to the total P content of dust. Why is this done? Also why are the values normalized to the values reported by Pett-Ridge et al.? Why not use the authors measured P content in the dust they applied? This approach needs to be clarified.
I am also confused by the units shown in Figure 4a. The units shown in the figure are mg m-2, but the figure title says P uptake per gr of dust. This adds to my confusion and needs to be clarified in the revised manuscript.
Lines 500- 505 Can the authors really measure these concentrations to hundredths of µg g-1 ? This seems too precise to me; too many significant figures.
Citation: https://doi.org/10.5194/egusphere-2026-4320-RC1 -
RC2: 'Comment on egusphere-2026-4320', Anonymous Referee #2, 15 Oct 2026
reply
This paper addresses an important topic, the long-range transport of nutrient elements (particularly phosphorus) that can occur whenever airborne dust is generated from terrestrial landscapes. Given increasing climatic extremes, this topic is very timely. The strength of this paper is its assessment of the fate of dust-derived P from Saharan sources once it is deposited in the tropical forest of Puerto Rico. The experiments presented are relatively novel, as they track the fate of dust applied in the field to two locally important tree species.
The discussion needs to be more focused. It wanders around and sometimes circles back to points that have already been made. For example, line 758 makes the point that internal recycling is important in the P cycle of this forest. The same point was already made as the topic sentence of an earlier paragraph (Line 736). Furthermore, the Discussion seems to calculate new data (line 762) rather than reflect on the data that have already been presented in the Results. Finally, the Discussion throughout assumes that total plant uptake from both external (dust, rain) and internal inputs such as leaf decomposition is the most important frame of reference. I disagree, however, and would argue that an equally important perspective is the overall balance of inputs to the forest ecosystem (deposition, weathering) and outputs (soluble and particulate P transported downstream in the case of this forest) that maintain the ecosystem reservoir of available P. Both perspectives are useful. Each should be clearly presented. I believe that the most effective and efficient way to do so would be to develop a figure that combines the external inputs of rain and dust, internal pools of soil and vegetation, internal transfers of litterfall and plant uptake, weathering rates, and hydrologic export.
Minor edits:
Throughout, I am unconvinced that ‘Fe anchored’ is the best terminology for describing this approach. Perhaps Fe-derived? Or calculated from Fe retention? The usage is unusual and is awkward for me.
Line 38: conservative anchor should be conservative tracer
Line 62: SWAP should be Swap
Line 121: a better reference is Murphy, S. F., R. F. Stallard, M. A. Scholl, G. González, and A. J. Torres-Sánchez. 2017. Reassessing rainfall in the Luquillo Mountains, Puerto Rico: Local and global ecohydrological implications. Plos One 12: e0180987.
Line 118: Consider expanding initial description to end with “Puerto Rico, which is a tropical broadleaf forest (Weaver and Murphy 1990)
Line 126: edit to “Precipitation in the forest is relatively aseasonal, with no month receiving less than 200 mm of rainfall; however, a drier period often occurs in February and March This point is already made explicitly in McDowell, W. H. and others 2012. Geographic and ecological setting of the Luquillo Mountains, p. 72-163. In N. Brokaw et al. [eds.], A Caribbean Forest Tapestry: The Multidimensional Nature of Disturbance and Response. Oxford University Press.
Line 268: exercised should be excised
Line 328 “gently brushed” should be gently brushed off
Line 382 Italicize species name
Line 461 and following seems to be methods with a tiny bit of results. Maybe the result should focus on how the different dust sources have similar or different responses in soil water and plants. Actual result is limited to line 474 et seq.
Line 539: no italics for Starr et al. 2023
Figure 3B check fonts for species names, they appear to be different
Fig. 4A and 4B remove figure headings; change 4A y axis legend as in Figure legend. 4B bioavailable is misspelled on y axis legend; Fig 5B Y axis Labile is misspelled.
Line 644: forests should be forest
Line 668: italicize species
Line 713 unlikely to be resolved should be unlikely to be measurable
Line 726: reference Gross et al. 2021, not Avner Gross
Line 760: words missing, does not make sense
Line 1172: authors should not be in all caps
Citation: https://doi.org/10.5194/egusphere-2026-4320-RC2 -
RC3: 'Comment on egusphere-2026-4320', Anonymous Referee #3, 20 Oct 2026
reply
This manuscript investigates how the dust deposition contributes to the P cycle in the humid tropical forest, which is limited in phosphorus. For this purpose, two experiments are proposed:
-1- a foliar "dust" application in the Luquillo Experimental forest
-2- a soil amendment of this forest by dust, Negev soils, and forest fire ash
The question of the fate of phosphorus in the dust deposited in these P-deficient ecosystems is very relevant and the first proposed experiment is very original.
The choice of the Luquillo Experimental forest site where dust deposits and the P cycle are well documented is also relevant.
The choice of materials applied to the leaves or to the soil is less relevant since these are not dusts deposited in Puerto Rico but original materials from Israel or transatlantic dust that are very different mineralogically (ie. high proportion of limestone) the dust deposits of Puerto Rico. Biogeochemical reactions can be very different.
The experimental protocol of experiment 1 has major flaws that will need to be addressed in order to interpret the results.
Even if protections (with large openings on the side) have been installed on the leaves, rain can certainly have preferentially entrained certain particles. Similarly, preferential mineral losses may occur at the time of particle application or collect. The absence of a mass balance (applied and collected particle mass) or of an overall chemical composition (Si, Al, Ca, Mg, K, Ti, Zr,...) for the various samples of collected particles does not allow to conclude whether the chemical evolution in Fe, P and forms of the measured P is related to a weathering of the particles or a preferential loss of certain dusts (for example, the shapes are rounded compared to the flattest ones or the largest particles).
Similarly, the quality of separation of particles and leaves is very sensitive in terms of interpretation of the results of chemical evolution and leaf. For example, if there are particles remaining on the leaf sample then the additional P (or Fe) is considered as foliar absorption. At the very least, an observation of control with a scanning electron microscope of the leaves after washing is necessary to validate this protocol.
In addition to the control of sampling, the calculation of P absorbed by the leaf also appears very problematic (ie Fe tracer, highly variable chemical concentrations of leaves,...): an isotopic tracer would seem more optimal.
In the absence of proof of the validity of this protocol, the conclusions on P transfers or shape changes are not supported, especially since the experiment only lasts on which days.
For these reasons, this manuscript cannot be published.
Moreover, even though the manuscript is generally well structured, it has defects such as the repetition of information in different paragraphs (2 to 5 times), the lack of precision in the material and method (ie the chemical and mineralogical composition of the Luquillo soil), the absence of legends and protocol (ie DRX) for additional data. I was surprised by the fact that organic matter was generally ignored throughout this study (contained in the materials used, in the mechanisms of release of P, etc.). The last two paragraphs of the discussion are interesting.
Citation: https://doi.org/10.5194/egusphere-2026-4320-RC3
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 241 | 290 | 35 | 566 | 41 | 38 | 32 |
- HTML: 241
- PDF: 290
- XML: 35
- Total: 566
- Supplement: 41
- BibTeX: 38
- EndNote: 32
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript (MS No.: egusphere-2026-4320) “Canopy processing of atmospheric dust creates a short-term phosphorus pathway in a tropical forest” by A. Gross et al. describes an interesting experimental study in which the authors (sort of) compare foliar uptake of phosphorus (P) from atmospheric particles with soil processing of P from atmospheric particles applied to soil. The authors are comparing apples with oranges but that notwithstanding, this is a challenging question to address. I like the paper and the authors should be commended for taking this research on. The authors provide a unique perspective on the important role of foliage in the processing and assimilation of atmospheric particles to obtain P. A challenge for the authors is to present a clear and coherent description of a complex experiment which contrasts foliar processing with soil processing of atmospheric particles using contrasting methods. Although I like the paper and think it should be ultimately published, I had much difficulty following the author’s text. I found myself going back and forth through the text trying to follow their experimental approach. I don’t think there is any way around this problem. The authors have undertaken a complex experiment and there are clear limitations, but the results are interesting and the paper could ultimately be published in EGU Sphere.
One of the big issues I have with the paper is understanding how the dust is “applied” to the leaves. The authors state that they “use a fine sieve to apply uniform coverage” But how is this done? The dust is a dry material. What keeps it from falling off the leaf after it is applied? Of the mass of dust applied to the leaves what fraction remains on the leaf? This experimental application should be clarified on p. 9, line 242.
Another concern that I have is that it seems that the mass of dust applied experimentally is quite large; approximately 70 g m-2 of leaf surface. The authors state that is approximately equivalent to 3 years of deposition. I understand that the authors would like to apply a large dose to observe a plant response, but could there be limitations associated with over dosing? If so these limitations should be discussed in the Discussion section of the paper.
I don’t like the fact that the authors made two separate applications of two different types of dust to the same leaves. I think this is a serious flaw in the experiment. How can the authors determine if one type of dust reacts substantially differently to the leaf than the other type. I understand that the authors had limited quantity of atmospheric dust, but this approach is a clear limitation to interpreting the experimental observations.
In various sections of the Discussion the authors address some of the limitations of the experiment. However, they don’t discuss the two limitations which I mention above and I consider these to be major limitations of the study. I think these limitations should be addressed in a separate “limitations” section of the Discussion which would include all the other limitations mentioned throughout the Discussion section.
I have a number of smaller points and comments which should be addressed before the manuscript is suitable for publication.
Line 133 I don’t understand “dry deposition is approximately 53 -73%”. Of what? Is this percent as P or total mass? If it is total mass, how is total mass expressed? Dry mass of wet plus dry deposition? The authors need to clarify what they are trying to say here.
Lines 177-180 How are the local dust and ash collected. Are these surface soil deposits? What depth? Some additional detail on the collection of these materials is needed. How soon after the fire was the ash collected?
Lines 261, 270, 379 and probably others elsewhere The authors cite Lokshin et al. 2026 and Lokshin et al. 2026a in the text. In the references they provide Lokshin et al. 2026a and 2026b. However this seems to be the same article cited twice. Something is incorrect and the authors should fix it.
Line 302 The authors should clarify here that the greenhouse experiment is Starr et al. 2023.
Line 393 It appears that the authors cite Figure 3a here is the text before they cite Figure 2. If I am correct the authors should probably switch the order of Figures 2 and 3.
Line 405 Should equation 3 be %ΔPest ?
Line 476 Why do the authors use ppm to express concentration units here when they use mg P g-1 elsewhere in the paper?
Line 484 The authors are not really normalizing the P fluxes to the soil surface area, are they? Aren’t the values normalized to the soil pot surface area (0.01 m2)? The soil presumably has a relatively high specific surface area and therefore a high surface area. This description should be clarified in the revised text.
Line 490 and Figure 4a I am completely confused by the normalization of the foliar P response to the total P content of dust. Why is this done? Also why are the values normalized to the values reported by Pett-Ridge et al.? Why not use the authors measured P content in the dust they applied? This approach needs to be clarified.
I am also confused by the units shown in Figure 4a. The units shown in the figure are mg m-2, but the figure title says P uptake per gr of dust. This adds to my confusion and needs to be clarified in the revised manuscript.
Lines 500- 505 Can the authors really measure these concentrations to hundredths of µg g-1 ? This seems too precise to me; too many significant figures.
Line 625 Here is the Discussion the authors first mention that the atmospheric acids could alter the availability of P in trans-Atlantic Saharan dust. It might be useful to mention this idea earlier in the Methods to justify using both desert soil dust and trans-Atlantic Saharan dust in the experimental design.