the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Drowning in a sandy ocean: Epiarenic growth of Tillandsia in the hyperarid Atacama Desert
Abstract. The Atacama Desert hosts a unique ecosystem formed by the sand-dwelling Tillandsia landbeckii, which extends over hundreds of square kilometers. This vegetation relies primarily on fog as its main water source; however, aeolian sand also plays a crucial role in the long-term persistence of both the species and the overall plant community. The terrain is sloped and exposed to the prevailing wind direction. Tillandsia forms regular banding patterns oriented orthogonally to these landscape features. In this study, we aim to elucidate the abiotic–biotic interactions between sand properties and vegetation characteristics through a comparative approach. Three populations – Caldera, Oyarbide and Arica –, each spanning several square kilometers in the southern, central, and northern regions of the Chilean Atacama Desert, were selected to compare wind regimes, terrain structure, sand and substrate properties, and vegetation structure in order to identify common principles that maintain vegetation integrity. Data were collected from six climate stations, 1,246 substrate samples, population genomic data from 718 individuals, as well as satellite imagery and digital terrain models. Our findings demonstrate that regional wind systems transport sand from distant source areas, while near the ground, Tillandsia vegetation reduces wind velocity and traps sand, leading to the formation of moderately sorted sandy substrates that are similar across all three populations. Sites lacking or containing dead Tillandsia individuals often differ significantly in substrate characteristics. Genetic analyses indicate that Tillandsia populations exhibit strong spatial structure albeit recruiting high genetic diversity and an excess of heterozygosity, reflecting adaptation to the dynamic environmental conditions. We conclude that sand represents an essential component of this ecosystem, while Tillandsia, as the dominant biotic factor, actively shapes and maintains this distinctive desert environment.
Status: open (until 22 Aug 2026)
- RC1: 'Comment on egusphere-2026-1429', Nicholas Lancaster, 22 May 2026 reply
-
CC1: 'Comment on egusphere-2026-1429', Xiaoping Yang, 29 May 2026
reply
The paper combing geomorphology, sedimentology, climatology and plant DNA offers an interesting perspective for experts in different fields, and particularly for those studying desert ecosystem both in Atacama and globally. The study shows that there is a correlation between grain size features, fog frequency and growth of plants in the Atacama. I am impressed by the efforts of collecting various kind of data including climate, DSM, sediments and plant DNA in such a remote area from Heidelberg.
I believe the paper could be better if the authors would consider some level of revisions. Also I think the section of Conclusions is too short. More details should be added to this section. I think the introduction section should be more focused. In its current version the introduction contains some of the general knowledge and it is not needed for a research paper. About five pages for introduction are indeed too long for my taste.
As the annual precipitation is extremely low in the study areas, it is needed to indicate the duration of the observations for calculating the mean annul precipitation in Fig. 1 and in the text.
As a geomorphologist I do not know much about methods of DNA. I believe the methodologies of sedimentology are fine. Even though I would like to ask about the meaning of sorting. See Lines 365-370: I think sorting degree cannot be directly measured with an absolute range due to general variations of grain size. Within coarse sands 1.27 um is a little but quite a lot if it is clay type of sediments.
Citation: https://doi.org/10.5194/egusphere-2026-1429-CC1 -
RC2: 'Comment on egusphere-2026-1429', Alessia Guggisberg, 16 Aug 2026
reply
General comments
This paper integrates a massive amount of granulometric, geomorphological, climatic and population genetic data from three different study sites, to understand the formation and maintenance of Tillandsia lomas, characteristic plant communities of the hyperarid Chilean Atacama Desert. Since grain size and index sorting differ between areas with or without Tillandsia, one may imply that the plants shape their own environment. Geomorphological and climatic data further indicate that banding patterns may only persist under a low wind energy system and are therefore highly sensitive to abrupt climate changes. The genetic results finally suggest an adaptive advantage of heretozygotes and emphasise the importance of clonal propagation for the establishment and long-term survival of Tillandsia. While previous studies have investigated the mechanistic and topographic characteristics of Tillandsia lomas, this is the first that specifically looks into the biotic effects on such a large scale.The authors provide an in-depth summary of their research results on Tillandsia, but fail to convey which questions remain unanswered. The paper reads a bit like a final report and it is not always clear, which data were newly produced for this paper and existed already from previous studies. Overall, the paper needs to be streamlined and better structured. Since this study integrates data from very different research areas (ranging from geomorphology to genomics), it is important to clearly formulate the questions or null hypotheses at the very beginning and explain why/how the methods that have been chosen may address those, so that a broad spectrum of researchers can grasp all the details of this study. Similarly, this paper could gain in significance, if (seemingly) incongruent/inconsistent results would be discussed more thoroughly. For example, the exposition analyses showed major differences between the study sites (Tillandsia at Caldera are not perfectly facing incoming winds and coppice dunes are more developed at Oyarbide), but the explanations for these differences lack clarity. Also, at lines 819-820, the authors finally state that wind and sand dynamic parameters are largely independent from landscape characteristics, but what this exactly means in light of the study questions remains open. Last but not least, it would be easier for the reader, if the data were presented and discussed in a similar order throughout the sections of the paper (e.g. first sand properties, then geomorphological aspects and finally genetic results).
Specific comments
More specifically, the Introduction must be shortened. Lines 80-87 seem out of scope, while paragraphs of lines 88-166 are too lengthy and read more like a textbook or a general report/review. The reader only gets introduced into the core of the topic at line 186 and into the knowledge gaps (or hypotheses) at line 220. Overall, one misses the exact reason for each type of analyses. For example, lines 605-609 clearly emphasise why grain size and sorting index got analysed, but such explanations come too late in the text and could be moved to the Methods section and be briefly mentioned in the Introduction when listing the goals of the study. The Methods are described in details, but it is not clear, whether the genetic data have been newly produced, or existed already and have simply been combined for this study. The Results and Figures nicely summarise the data, but the former do not refer to the (two) hypotheses posited in the Introduction. Some of the sentences would also better fit as arguments in the Methods or as explanations in the Discussion. The Discussion would benefit from a clearer structure that mirrors the questions and goals posited in the Introduction. The Conclusion is too short and general, and does not do justice to the significance of the results. The authors briefly mention that tillandsiales are threatened by climate change, but they do not share their opinion on how the community may evolve, if the conditions would remain stable or would further change over time, which of the study site may be more resilient, or which of the parameters (sand/wind dynamics, genetics, etc.) are most relevant for the plants to survive. Accordingly, the Abstract and Short Summary can be improved, to better emphasise the knowledge gaps, goals and methods used to address these, and increase the relevance of the study for researchers working on desert ecosystems.
Technical comments
L. 29: …Tillandsia landbeckii known as Tillandsia lomas or tillandsiales, which…
L. 32: which other plant community?
L. 29-35: One the one hand, you certify that sand plays a crucial role in the persistence of the plant community, but on the other hand this is exactly what you want to investigate. Could you better phrase what one still doesn’t know, what you exactly want to test?
L. 46: Could you clarify which substrate characteristics change the most?
L. 46-49: Based on the Abstract, the reader cannot grasp the relevance of the genetic results.
L. 50: The 1st sentence is weak; we know that sand plays a central role since line 31. Could you rather conclude on how the dynamics may evolve with the prevailing conditions and what would be the long-lasting consequences?
L. 71: What do you mean by ‘predominant grain sizes’?
L. 72-75: Are the two statements saying the same, namely that heterozygous genotypes are fitter?
L. 88: What is the difference between plant communities and vegetation units?
L. 88: …forming plant communities and vegetation units, whereby biotic and abiotic…
L. 88: new paragraph after (Willi et al., 2002).
L. 139: …monospecific stands of shrubs and perennial grasses
L. 142: simply altered or rather enhanced fire regimes?
L. 146: …these interactions cause noticeable spatial structures in monospecific vegetation.
L. 152: …vegetation patterns, which are finally reduced to highly isolated patches…
L. 154: …formed by (various sets) of plant species…
L. 155: new paragraph after “…across continents”.
L. 163: …contour lines, which occur in Africa…
L. 190: …”Tillandisa lomas” (or tillandisales)
L. 197: new paragraph after (Rundel et al., 1997)
L. 214: …vegetative growth strategy for this long-lived perennial plant.
L. 226-227: What do you mean here, that if wind/fog conditions change over time, we would observe a concerted response of the community over time?
L. 232: To verify these hypotheses, we investigated three…
L. 234: …using population genomic data with…
L. 245: punctuation or words missing in the sentence?!
L. 258: …Chilean populations herein named Arica (Fig. 1).
L. 262: …characterized by well-developed…
L. 269: ...according to results from genetic assignments.
L. 276: …Humboldt Current, whereby coastal upwelling…
L. 280: …daily sea-breeze…
L. 287: From a geological and a geomorphological point of view, the three populations experience different settings.
L. 291: …as possible sediment sources.
L. 292-293: …sand/silt matrix are scarce.
L. 294: …are most probably providing the source…
L. 296: …sand and silts are dominant.
L. 312: …we gathered data on…
L. 326: How much ‘in front’, did you fix a given distance?
L. 353-354: …filled two-third with dH2O.
L. 412-427: Does it mean you combined data from three different sequencers?
L. 445-446: ….across all GBS experiments.
L. 449: A dataset without duplicates…
L. 452-454: In addition, genetic assignment tests were also performed for all three study fields to ensure that historical genetic differentiation processes are not biasing the analyses.
L. 482: Did the year of the measurements (March 2022 – March 2023; see line 482) fit within the average annual fluctuations or should it rather be considered as exceptional (e.g. Niño year)?
L. 504: new paragraph after ‘on site’. Has the recording at Oyarbide been done during the same period (March 2022-March 2023) as at Caldera and Arica?
L. 536: The total vector units were accumulated…
L. 558: …is about 14 cm and the data are freely available.
L. 584-587: rephrase (incomprehensible)
L. 652: …size from Caldera (Fig. 2c, d).
L. 653: …the coarsest grain size and Oyarbide the fines…
L. 658-659: Since the three sites show different results regarding grain size and sorting index, can we actually conclude with a general trend as in header of §3.1? And are the observed differences between sites necessarily the result of different aeolian sand origins?
L. 660: How do you know the sands came from distant areas? see also statement at line 43 in the Abstract and line 904 in the Discussion
L. 667-668: What biases could therefore be introduced, how could this error rate affect the results and how trustful are then the conclusions drawn from these analyses?
L. 672-673: …comparisons across all sites with past analyses…
L. 677: …not attributed to increased outbreeding but rather preferential clonal propagation…
L. 724: Could you formulate a summary sentence for each header, as for §3.1 and §3.2?
L. 728: Study fields at Arica and Caldera…
L. 743: …channelled through the valley, from which no wind data records exist.
L. 766: Is there an error in the sentence, like a word missing?
L. 768: …direction (Caldera).
L. 775: …and Caldera (MSC3 and 1).
L. 786: Do you mean ‘aspects’ by ‘exposition’?
L. 791-793: Do you mean the dunes at Oyarbide are so high that on the lee side, no Tillandsia can grow? Please, reformulate more clearly.
L. 815: …the inclined plane is slightly rotated.
L. 813-816: I do not get this point!
L. 818: ‘Between the two groups’, do you mean sand/wind vs. terrain parameters? Please, specify.
L. 827: Why shall the tillandsiales be confusing and illogical?
L. 833: Specify the word ‘that’ at the beginning of the sentence.
L. 835-839: redundant sentences
L. 859-862: Any explanation for the differences between the lower and upper part of the study site?
L. 885: ‘…they transport’ shouldn’t refer to both wind and sand, I think, but only wind.
L. 891: What are these ‘correlating patterns’?
L. 909 & 917: ‘Tillandsia’ should be italicised.
L. 935-937: …in the phyllosphere of Tillandsia have been proposed as potential source of nitrogen, which may have a direct effect on growth and need to be fine-tuned…
L. 988: missing reference for the dieback
Graphical abstract: I could not fully grasp every graphic. To me, the 1st lane represents the banded pattern of Tillandsia lomas at the three study sites and highlights the analyses on grain size and index sorting. The 2nd lane emphasises the different sand and wind dynamics in grids occupied by dead/living plants or devoid of vegetation. Logically, the 3rd lane should present the genetic results, but I do not get the message.
Fig. 7: Add site information in the legend or graph.
Fig. 8: Could you make the graphs readable for red-green colour-blind readers?Citation: https://doi.org/10.5194/egusphere-2026-1429-RC2
Viewed
Since the preprint corresponding to this journal article was posted outside of Copernicus Publications, the preprint-related metrics are limited to HTML views.
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 309 | 0 | 2 | 311 | 0 | 0 |
- HTML: 309
- PDF: 0
- XML: 2
- Total: 311
- BibTeX: 0
- EndNote: 0
Viewed (geographical distribution)
Since the preprint corresponding to this journal article was posted outside of Copernicus Publications, the preprint-related metrics are limited to HTML views.
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This is an interesting manuscript that combines geomorphology and DNA-based population genetics to understand monospecific populations of Tillandsia in the hyper arid Atacama Desert and their relations to sand transport by wind.
The manuscript is very long and reads like a report to sponsors. It could benefit from severe pruning to make it more concise and to tell the authors' story more effectively.. At the same time, it does lack information on the physical and biotic character of the investigated sites. Numerous correlations between variables are discussed but no plots are included.
Introduction – very long and introduces multiple topics – monospecific plant communities, dryland ecosystems, banded vegetation patterns, Tillandsia communities. We finally learn the objectives of this paper at line 220.
Theis section needs to focus on Tillandsia ecosystems in the Atacama and their distinctive characteristics and not ecological theory. Some subheadings would be helpful
Methods
Very detailed documentation of sample collection and analysis.
Results
This section could benefit from some reorganization of material in order to present first the physical environment and then the biotic information. This section could benefit from some site photographs to show the environments in which the Tillandsia grow
785 – it is aspect, not exposition here and in this section
Discussion
This section is very generalized and discusses broader aspects of Tillandia ecology especially in relation to climate change. It does not easily follow from the result presented here. Some more detailed discussion of relationships between topography, winds, sand transport and particle size should be included here
Conclusions
do not follow from the results of the field and laboratory studies documented here