the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 30 Jan 2026
Unlocking the air: DNA metabarcoding sheds light on seasonal fungal dynamics in a temperate floodplain forest
Abstract. Airborne fungi play a pivotal role in ecosystem functioning, agriculture, people health and wellbeing, yet their response to increasingly frequent climate extremes remains poorly understood. There is thus a need for long-term studies that can capture both seasonal and annual dynamics of fungal remnants in the air. Here, we applied DNA metabarcoding of the ITS2 region to investigate the composition and responses of fungal aerosols to meteorological variables in a temperate floodplain forest habitat. Passive air samples were collected continuously at three heights above the ground between March 2019 and February 2020 at the Leipzig Canopy Crane (Germany). Fungal aerosol assemblages were found to be dominated by Ascomycota (74.3 %) and Basidiomycota (25.1 %), with the genera Cladosporium, Epicoccum, and Alternaria consistently prevailing across samples. Our results revealed that seasonal changes in air temperature were the primary driver for compositional changes in fungal aerosols, with Ascomycota increasing in abundance during warmer months and Basidiomycota dominating during colder months. Through abundance differential analysis, we identified 66 genera, including allergenic and pathogenic taxa, that shifted significantly in abundance with seasonal temperatures. Interestingly, neither sampling height nor humidity had a significant effect. Our study highlights the importance of conducting long-term monitoring of bioaerosols under changing climate conditions while also creating a benchmark for future comparative studies.
- Preprint
(4782 KB) - Metadata XML
-
Supplement
(1750 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-372', Anonymous Referee #1, 02 Mar 2026
-
AC1: 'Reply on RC1', Beatriz Sánchez-Parra, 02 Apr 2026
General comment: We would like to thank the reviewer for taking the time to review our manuscript, and for their thoughtful and constructive comments. Their feedback has been invaluable in enhancing the clarity, rigour and overall quality of our work.
Below, we provide detailed, point-by-point responses to each of the reviewer’s comments.
We have also uploaded the changed figure, and supplementary files as attached documents.
Specific comments
Materials and methods. The criteria and distance between the tree gaps are not mentioned (line 127). This may be relevant to support or discuss the lack of differences.
We thank the reviewer for highlighting this detail. The distance between the sampling stations (locations), which is now included in the article (Materials and methods section and as new supplementary figure 3.3), was only about 40 to 60 metres. Since these distances are not very great—unlike, for example, distances measured in kilometers—this could be one of the reasons why no differences were observed. We address this point in the Discussion section.
Three different sets of primers were employed for NGS libraries. Is a novel approach? If not, please cite previous works. It is not specified whether the sequences were or not pooled. I assume they were, but please clarify. Also, if this is a novel approach, the differences found in richness or diversity between them should be mentioned and a recommendation for the specific set of primers yielding the best resolution.
We thank the reviewer for this comment. The use of multiple primer pairs in metabarcoding studies is not novel and has been previously proposed to improve taxonomic coverage (e.g. Tedersoo et al., 2015). In our study, three primer pairs were selected based on this work to maximise the detection of fungal diversity while maintaining cost efficiency. All PCR reactions were performed on the same DNA extract, and the resulting amplicons were pooled prior to sequencing. Because amplicons from different primer pairs were not individually barcoded, it was not possible to assess or compare the performance of each primer combination in terms of richness or diversity. Therefore, no primer-specific comparisons or recommendations can be made for our samples. We have clarified these points in the Methods section, including the papers that have previously demonstrated the efficiency of the primers used here.
For the altitudes employed, I doubt about the relevance of the mass air trajectories and the accuracy, especially within a forest. Please, discuss and justify.
We thank the reviewer for this comment. We agree that, at the sampling heights considered and within a forested environment, the accuracy and relevance of backward air mass trajectories are limited, and local processes are likely to dominate. In our study, trajectory analysis was used only as a qualitative assessment to verify whether sampling heights were exposed to comparable air masses, and thus to support the decision to merge samples across heights. It was not used to infer precise source regions or transport processes. We have clarified this point in the Discussion.
Results. A figure or table indicating the taxa of the core mycobiome would be appreciated and relevant for future comparable studies.
We thank the reviewer for this comment. We have now added Table S3.4.
Figure 3 can be improved (legend names are cut and units are not necessary).
Done, legend names are written in full and units have been removed.
A statistical test must be added as support for Figure 4 and statement in lines 273-275.
We thank the reviewer for this comment. We have now added to the text: “Differences in relative abundance between warm and cold months were assessed using a Wilcoxon rank-sum test. Ascomycota showed significantly higher relative abundance during the warm season, whereas Basidiomycota was significantly more abundant during colder months (Wilcoxon test, p = 6.68 × 10⁻⁵ and p = 2.0 × 10⁻⁶, respectively).”
Line 283. …” in several other studies”. Discuss the relevance (similar or different environment?). In addition, this and many lines in page 11 are more suitable for the Discussion section.
We thank the reviewer for their comment, we have now moved these lines to the Discussion and clarified this point in this way: “In line with this, the dominance of genera such as Cladosporium, Alternaria, and Epicoccum observed in this study is consistent with previous works conducted in temperate environments, where these taxa are commonly reported as major components of airborne fungal communities (Almaguer et al. 2014; Sadyś et al. 2015; Akgül et al. 2016; Grinn-Gofroń et al. 2018; Antón et al. 2019; Ščevková and Kováč 2019; Grinn-Gofroń et al. 2020). These three genera are also some of the best-known sources of allergic reactions and plant diseases (Bavbek et al. 2006; Abuley and Nielsen 2017; Nowakowska et al. 2019; Grinn-Gofroń et al. 2020). Their increased abundance during the summer months supports the hypothesis that Ascomycota are favoured under warmer conditions, likely reflecting enhanced sporulation and dispersal during these periods.
In contrast, members of Basidiomycota, particularly within the class Tremellomycetes, were more prevalent during colder periods, indicating a seasonal pattern opposite to that observed for Ascomycota. The detection of genera such as Vishniacozyma and Itersonilia, both associated with plants as endophytes or pathogens (Liu et al., 2025; Gandy, 1966), further supports the role of surrounding vegetation as a source of airborne fungal propagules.
At a finer taxonomic resolution, the species Vishniacozyma tephrensis, characterised as psychrotolerant and psychrophilic, was predominantly detected during colder periods, with higher abundance in late autumn. Its occurrence under these conditions reflects its tolerance to low temperatures and environmental stress (Vishniac 2002; Wei et al., 2022). In this context, it is worth noting that, while environmental factors such as temperature, humidity, precipitation, and wind speed…”
Discussion. Paragraph 408-429 can be shorter.
Done, the paragraph has been revised and shortened to improve clarity and readability, while adding key ecological interpretations to address reviewer’s 2 comments on seasonality. The paragraph now reads as follow: “In this context, the detection in our dataset of the plant pathogen Erysiphe alphitoides, a widespread pathogen causing oak powdery mildew in Europe (Marçais et al. 2014; Tăut et al. 2024), is consistent with the presence of suitable host plants (Quercus spp.) and likely reflects seasonal patterns in host phenology and canopy development (Andrew et al. 2017). More broadly, the occurrence of multiple phytopathogenic taxa, including species such as Zymoseptoria tritici, which affects wheat plantations (Torriani et al., 2015), suggests that surrounding vegetation and nearby agricultural areas represent potential sources of airborne fungal propagules. These observations indicate that seasonal changes in plant communities and phenology play a key role in shaping airborne fungal assemblages, alongside the influence of meteorological variables. In addition, it is important to note that climate change may further modulate these patterns by altering host plant phenology (Piao et al., 2019). In this context, we argue that disentangling the effects of meteorological conditions from broader seasonal dynamics remains challenging, as previously reported (Cáliz et al. 2018; Bowers et al. 2012), and that more robust assessments of these drivers require multi-year sampling.
Interestingly, the detection of taxa that have been previously reported primarily outside Europe e.g. Alanphillipsia aloetica from South Africa (Crous et al., 2013), further highlights the complexity of airborne fungal communities and the potential for long-distance dispersal. Taken together, these findings provide a baseline for assessing temporal dynamics and the impacts of environmental change in airborne fungal assemblages, thereby corroborating the role of this study as an important benchmark for future works.”
Lines134-139 are confusing. Please simplify.
We thank the reviewer for this comment, this has now been simplified to: “For this study, we analysed one week of samples for each month between March 2019 and February 2020 (Table S2.2). Specifically, the 4th week of each month was selected due to logistical and financial constraints and to ensure comparability with meteorological records across all months”.
Line 96. … “we would expect?”
This has now been changed with: “only limited divergence in the composition of fungal aerosols collected within the canopy layer is expected.”
Lines 334, 335: italics for the cited genera.
Done.
Figure S2. A mix of languages in the figures
We thank the reviewer for this comment, this was changed to English.
- AC2: 'Reply on RC1', Beatriz Sánchez-Parra, 02 Apr 2026
-
AC1: 'Reply on RC1', Beatriz Sánchez-Parra, 02 Apr 2026
-
RC2: 'Comment on egusphere-2026-372', Anonymous Referee #2, 12 Mar 2026
General comment:
The authors investigate fungal diversity in air samples collected at three locations and heights over the course of one year, including both summer and winter sampling, in a forest environment. In the introduction, the authors note that bioaerosol composition depends on the emission source, and therefore seasonal variation is expected. The manuscript places considerable emphasis on correlations with meteorological parameters. At the same time, the sampling design is likely influenced by seasonal changes related to tree phenology, vegetation dynamics, and fungal life cycles. These broader seasonal processes are very likely reflected in the results, with temperature acting as one contributing factor. The authors also acknowledge this in the discussion (lines 352–361), where they state that environmental factors influence the phenology of source species and thereby the spore‑forming cycles of fungi. As a result, the findings appear to primarily reflect seasonal changes, which is also consistent with the manuscript’s proposed title. In this context, a clearer focus on seasonal variation could strengthen the overall narrative of the study.
Many phytopathogens are present in the dataset, and it is therefore expected that these taxa peak when host plants and canopy are present. Similarly, it is reasonable that these taxa occur predominantly during the warmer months, while wood‑ and plant‑debris–decaying fungi peak later in the season during cooler periods. Given the current design, it may be challenging to disentangle the direct effects of meteorological variables from the broader seasonal context. A more robust assessment of direct meteorological influences would likely require multi‑year sampling.
The dataset includes many phytopathogens and parasitic fungi, which understandably show seasonal patterns. However, the manuscript would benefit from deeper reflection on the identities and ecological relevance of the detected species. Some taxa appear to have been previously reported only outside Europe (e.g., Alanphillipsia aloetica from South Africa). Conversely, oak powdery mildew (Eryisphe alphitoides) is described as an invasive species, although to my knowledge it is already widespread across Europe and present on multiple continents. The detection of Zymoseptoria tritici, a wheat pathogen, in some samples (Table S3) likely reflects nearby wheat fields as a source and further supports the interpretation that the dataset captures seasonal dynamics.
In my view, the stated aim of providing a foundation for studying the impact of climate change on ecosystem functioning, agriculture, and human health is not fully achieved in the current version. However, with revisions that align the manuscript more closely with the seasonal patterns clearly present in the data, and with consideration of the points raised above, the study has the potential to make a valuable contribution and would be suitable for publication.
Comments on materials and methods
The authors state that the ITS2 region was used for the study; however, the primers listed in the Materials and Methods section will also amplify ITS1. It is unclear why the authors amplified the entire ITS region and assembled sequences across this region if only ITS2 was used in subsequent analyses. It is also unclear whether all analyses were based on assembled sequence data or whether assembly was used only for taxonomic identification. Please clarify this. The geographical location of the sampling stations is not described in sufficient detail. The methods state that “three distinct tree gaps were identified on site and designated as sampling stations”— How far into the forest were they located and how far apart were these gaps, and were the samples compared individually or pooled? It is also unclear how the weekly samples were selected. Was only one week per month chosen for analysis? According to Table S2, sampling appears to have occurred approximately every fourth week. This needs to be detailed in the materials and methods section, although listed in Table S2
Technical details
Line 145-152: Please list the primer pairs in a table and indicate which ITS region they are amplifying
Line 188: How was the data normalized?
Line 233: To stay consistent “areas” should probably be changed to “sources”
Line 218: it says that data is grouped into 760 ASVs, according to Table S3 1,877 ASVs were identified (Table S3). Please explain this discrepancy.
Table S3: Column (sample?) labeling should be clarified, month and/or dates should be added. Heights and locations should be explained. It is unclear why some entries are highlighted or colored.
In Figure 2, it is unclear why the correlation between wind speed and Shannon diversity is not significant, while the correlation with Gini–Simpson diversity is significant. Shannon and Gini–Simpson indices are highly correlated. This may need some discussion.
Figure S1: what does the different colours of the trajectories indicate? Explanation is needed
Citation: https://doi.org/10.5194/egusphere-2026-372-RC2 -
AC3: 'Reply on RC2', Beatriz Sánchez-Parra, 02 Apr 2026
General comment: We thank the reviewer for this thoughtful and constructive comment. We agree that seasonal dynamics, including vegetation phenology and fungal life cycles, are likely to play a major role in shaping airborne fungal communities, alongside meteorological variables. In response, we have revised the Discussion to more explicitly acknowledge the contribution of seasonal processes and to clarify that the observed patterns likely reflect an interplay between meteorological conditions and broader seasonal dynamics. We have also expanded the ecological interpretation of key taxa. In particular, we now discuss the occurrence of phytopathogenic species such as Erysiphe alphitoides in relation to host plant phenology, and interpret the detection of Zymoseptoria tritici as likely reflecting nearby agricultural sources. Additionally, we have moderated statements regarding taxa with uncertain biogeographic distributions (e.g. Alanphillipsia aloetica) by emphasising the potential role of long-distance dispersal and limitations in species-level taxonomic resolution.
Finally, we now explicitly acknowledge that disentangling the direct effects of meteorological variables from broader seasonal processes remains challenging within the scope of a single-year study, and that multi-year sampling would be required for a more robust assessment. These revisions align the manuscript more closely with the seasonal patterns evident in the data, while maintaining the relevance of meteorological drivers.
We provide detailed, point-by-point responses to each of the reviewer’s comments.
We have also uploaded the supplementary files with the modifications made as attached documents.
Line 145-152: Please list the primer pairs in a table and indicate which ITS region they are amplifying
We thank the reviewer for raising this point. We have clarified in the Material and methods section that multiple primer pairs targeting different portions of the fungal ITS region were used. As a result, the dataset includes sequences from both ITS1 and ITS2 regions, and we have revised the text accordingly to refer more generally to the fungal ITS region. The text reads now as: “The fungal internal transcribed spacer (ITS) region was amplified using three different primer combinations (Table 1). These were selected based on previous work aiming to maximise fungal taxonomic coverage (Tedersoo et al., 2015; White et al., 1990). PCR reactions were performed on the same DNA extract, and the resulting amplicons were pooled prior to sequencing (see below for specific details). Because amplicons from different primer pairs were not individually barcoded, it was not possible to assess primer-specific differences in richness or diversity.”
We have also added a table listing all primer pairs used and indicating the corresponding fungal ITS region amplified by each pair (Table 1).
This is the table:
ITS1Fngs – ITS2
GGTCATTTAGAGGAAGTAA
GCTGCGTTCTTCATCGATGC
ITS1
Tedersoo et al. 2015; White et al. 1990
ITS3 – ITS4
CATCGATGAAGAACGCAG
TTCCTCCGCTTATTGATATGC
ITS2
White et al. 1990
ITS1ngs – ITS2
TCCGTAGGTGAACCTGC
GCTGCGTTCTTCATCGATGC
ITS1
Tedersoo et al. 2015; White et al. 1990
While the use of multiple primer pairs may introduce some amplification bias, this does not affect the main ecological patterns reported, which are based on relative differences in community composition across samples.
Line 188: How was the data normalized?
We thank the reviewer for this comment. We have clarified that normalisation was performed using total-sum scaling (relative abundance transformation), whereby ASV counts were divided by the total number of reads per sample and expressed as percentages.
Line 233: To stay consistent “areas” should probably be changed to “sources”
Thank you, this has now been changed to “sources”.
Line 218: it says that data is grouped into 760 ASVs, according to Table S3 1,877 ASVs were identified (Table S3). Please explain this discrepancy.
We thank the reviewer for pointing out this apparent discrepancy. Table S3 reported the total number of ASVs identified after removal of contaminants (1,877 ASVs). In contrast, the 760 ASVs reported in line 218 correspond to the subset retained after additional filtering steps applied prior to downstream analyses. Specifically, ASVs with low abundance (≤2 reads) and low prevalence (present in less than 1% of samples) were removed to reduce noise and improve comparability among samples. This filtering step substantially reduced the number of ASVs while retaining the vast majority of sequence reads.
We have clarified this distinction in the revised manuscript. New text: “Our quality-filtered dataset consisted of 1,421,177 fungal sequencing reads, with an average of 13,198 reads per sample. After removal of contaminants, 1,877 ASVs were identified. Following additional filtering of low-abundance and low-prevalence ASVs, 760 ASVs were retained for downstream analyses (Table S3.1).” and we have also modified the Table S3.1 accordingly.
Table S3: Column (sample?) labeling should be clarified, month and/or dates should be added. Heights and locations should be explained. It is unclear why some entries are highlighted or colored.
We thank the reviewer for this comment. We have clarified the sample naming convention in the caption of Table S3. Sample identifiers now follow the format tXhY_YYMM, where tX indicates the sampling station location, hY the sampling height (3 m, 15 m, and 28 m), and YYMM the last two digits of the year and the month of sampling. An example (t1h1_1903) has been included in the caption of the table for clarity. Conditional formatting has been disabled from the entire sheet.
In Figure 2, it is unclear why the correlation between wind speed and Shannon diversity is not significant, while the correlation with Gini–Simpson diversity is significant. Shannon and Gini–Simpson indices are highly correlated. This may need some discussion.
We thank the reviewer for this comment. We agree that Shannon and Gini–Simpson indices are strongly related and, in our dataset, both showed a positive correlation with wind speed. However, the two tests yielded slightly different significance levels: Shannon (Spearman’s ρ = 0.56, p = 0.058) and Gini–Simpson (Spearman’s ρ = 0.61, p = 0.034). We therefore interpret this as a borderline difference around the significance threshold rather than a biological contradiction. Because the analyses were conducted on monthly pooled data, the sample size was limited, and small differences in index formulation and rank structure were sufficient to place one result just above and the other just below p = 0.05. This has been clarified in the main text.
Figure S1: what does the different colours of the trajectories indicate? Explanation is needed
We thank the reviewer for this comment. We have clarified in the caption of Figure S1 that colours represent trajectories corresponding to different days of the sampling period. -
AC4: 'Reply on RC2', Beatriz Sánchez-Parra, 02 Apr 2026
When submitting the first reply, we realized that Table 1 had not been submitted correctly in the previous comments. Here it is correctly submitted, with the table description included.
Table 1. Primer pairs used for ITS amplification, including primer sequences, target region, and references.
ITS1Fngs – ITS2
GGTCATTTAGAGGAAGTAA
GCTGCGTTCTTCATCGATGC
ITS1
Tedersoo et al. 2015; White et al. 1990
ITS3 – ITS4
CATCGATGAAGAACGCAG
TTCCTCCGCTTATTGATATGC
ITS2
White et al. 1990
ITS1ngs – ITS2
TCCGTAGGTGAACCTGC
GCTGCGTTCTTCATCGATGC
ITS1
Tedersoo et al. 2015; White et al. 1990
Citation: https://doi.org/10.5194/egusphere-2026-372-AC4 -
AC5: 'Reply on RC2', Beatriz Sánchez-Parra, 02 Apr 2026
Again, we're sorry, as the table 1 wasn't sent correctly in the second reply either.
This would be the good one:
Primer pair
Forward primer (5'-3')
Reverse primer (5'-3')
Target region
Reference
ITS1Fngs – ITS2
GGTCATTTAGAGGAAGTAA
GCTGCGTTCTTCATCGATGC
ITS1
Tedersoo et al. 2015; White et al. 1990
ITS3 – ITS4
CATCGATGAAGAACGCAG
TTCCTCCGCTTATTGATATGC
ITS2
White et al. 1990
ITS1ngs – ITS2
TCCGTAGGTGAACCTGC
GCTGCGTTCTTCATCGATGC
ITS1
Tedersoo et al. 2015; White et al. 1990
Citation: https://doi.org/10.5194/egusphere-2026-372-AC5
-
AC3: 'Reply on RC2', Beatriz Sánchez-Parra, 02 Apr 2026
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 248 | 101 | 32 | 381 | 25 | 17 | 34 |
- HTML: 248
- PDF: 101
- XML: 32
- Total: 381
- Supplement: 25
- BibTeX: 17
- EndNote: 34
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
This study addresses an interesting and relevant topic, contributing to the understanding of how environmental parameters shape bioaerosol composition within a specific ecosystem. Although their findings are generally consistent with previous studies, there remains a scarcity of data from environments such as the one investigated here, particularly studies providing long-term temporal resolution (one year). The scientific rationale motivating this work is comprehensively addressed; however, certain explanations lack clarity, and the organization of ideas could be improved in some sections of the manuscript.
The methodology employed—NGS-based sequencing—is appropriate for this type of study. In addition, the multi-marker metabarcoding strategy is commendable, as it attempts to mitigate primer-related biases. The statistical analyses are properly conducted, and the manuscript is, overall, well written and easy to follow. Nevertheless, several aspects could benefit from further refinement (see detailed comments below).
Globally, I find this work suitable for publication when addressing the following changes.
Specific comments
Materials and methods. The criteria and distance between the tree gaps are not mentioned (line 127). This may be relevant to support or discuss the lack of differences.
Three different sets of primers were employed for NGS libraries. Is a novel approach? If not, please cite previous works. It is not specified whether the sequences were or not pooled. I assume they were, but please clarify. Also, if this is a novel approach, the differences found in richness or diversity between them should be mentioned and a recommendation for the specific set of primers yielding the best resolution.
For the altitudes employed, I doubt about the relevance of the mass air trajectories and the accuracy, especially within a forest. Please, discuss and justify.
Results. A figure or table indicating the taxa of the core mycobiome would be appreciated and relevant for future comparable studies.
Figure 3 can be improved (legend names are cut and units are not necessary).
A statistical test must be added as support for Figure 4 and statement in lines 273-275.
Line 283. …” in several other studies”. Discuss the relevance (similar or different environment?). In addition, this and many lines in page 11 are more suitable for the Discussion section.
Discussion. Paragraph 408-429 can be shorter.
Technical corrections
Lines134-139 are confusing. Please simplify.
Line 96. … “we would expect?”
Lines 334, 335: italics for the cited genera.
Figure S2. A mix of languages in the figures