Accepted Manuscript

High similarity in bacterial bioaerosol compositions between the free troposphere and atmospheric depositions collected at high-elevation mountains

Xavier Triadó-Margarit, Joan Caliz, Isabel Reche, Emilio O. Casamayor

PII: S1352-2310(19)30068-8 DOI: https://doi.org/10.1016/j.atmosenv.2019.01.038 Reference: AEA 16527

To appear in: Atmospheric Environment

Received Date: 1 September 2018 Revised Date: 26 December 2018 Accepted Date: 27 January 2019

Please cite this article as: Triadó-Margarit, X., Caliz, J., Reche, I., Casamayor, E.O., High similarity in bacterial bioaerosol compositions between the free troposphere and atmospheric depositions collected at high-elevation mountains, Atmospheric Environment, https://doi.org/10.1016/j.atmosenv.2019.01.038.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT 1 High similarity in bacterial bioaerosol compositions between the free 2 troposphere and atmospheric depositions collected at high-elevation 3 mountains 4 5 Xavier Triadó-Margarit $1, Joan Caliz $1, Isabel Reche 2 and Emilio O. Casamayor* 1 6 7 1Integrative Freshwater Ecology Group, Centre of Advanced Studies of Blanes-Spanish Council 8 for Research CEAB-CSIC, Blanes E-17300, Spain; 2Departamento de Ecología and Instituto del 9 Agua, Facultad de Ciencias, Universidad de Granada, 18071 Granada, Spain 10 11 *Corresponding author. E.O. Casamayor. CEAB-CSIC, Accés Cala St Francesc, 14. E-17300. 12 Blanes (SPAIN). Phone: +34 972 336 101, Fax: +34 972 337 806. Corresponding author E- 13 mail: [email protected] 14 15 $equal contribution 16 17 Highlights 18 - Above the boundary layer ground-sampled MANUSCRIPTbacterial bioaerosols were largely 19 convergent with those from the free troposphere, both in composition and structure 20 - Bioaerosols passively deposited at high-elevation mountains are a good proxy for long- 21 term monitoring of the high atmosphere long-range transported microbiome 22 23 24 25 Keywords 26 Bioaerosol; aerosol deposition; free troposphere; above boundary layer, high-elevation 27 mountain; bacteria; 16S rRNA gene 28 29 ACCEPTED 30 Running title 31 Atmospheric deposition mimics free troposphere bacterial bioaerosols 32 33 34

1 ACCEPTED MANUSCRIPT 35 36 Abstract 37 The long-range transport of bioaerosols takes place in the free troposphere and has lately gained 38 a renewed interest in both environmental and health-related disciplines. Sampling free 39 troposphere bioaerosols has been, however, historically challenging and requires of expensive 40 and complex facilities. We analysed different bacterial bioaerosols studies carried out by 41 sequencing of the 16S rRNA gene, available from the literature. The dataset was compared with 42 bacterial bioaerosols present in rain and dry deposition passively collected at high-elevation 43 sites in Sierra Nevada along a set of sampling periods lasting 3 years. Up to 65 % of OTUs and 44 82 % of the bacterial genera were shared between wet and dry bioaerosols. Interestingly, only 45 Oxalobacteraceae were notably more abundant in wet deposition, with Noviherbaspirillum and 46 Massilia as dominant genera. We demonstrated that the bacterial composition of bioaerosols 47 collected by passive natural deposition at high-elevated mountains were closer to the bacterial 48 microbiome from the free troposphere. Interestingly, the meta-analysis showed a different 49 bacterial composition and community structure in bacterial bioaerosols collected at low- 50 elevated areas, over the open ocean, or during desert dust events. Since the boundary layer can 51 be easily reached in high mountain areas, and the local landscape is surrounded by rocks and 52 meadows, alpine stations are potentially optimal research sites with reduced influence of surface 53 aerosols, minimizing local contaminations. ConsequeMANUSCRIPTntly, sampling alpine bioaerosols could be 54 a good proxy for bioaerosols monitoring, long-range dispersal studies, and the dynamic 55 characterization of the free troposphere microbiome. 56 57 1. Introduction 58 Global-scale bioaerosols dispersal is an ubiquitous phenomenon (Griffin et al., 2017). 59 Bioaerosols originated from different terrestrial and marine sources can reach the free 60 troposphere and experience long-range intercontinental transport. Depending on wind 61 conditions and particles size, bioaerosols can be injected above the boundary layer at typical 62 heights from 500 m to 2 km before being washed out by wet (rain and snow) or dry deposition 63 (Burrows et al., 2009). Due to their microscopic size, bioaerosols can remain airborne and 64 viable for longACCEPTED periods (Barberán et al., 2014; Hervàs et al., 2009) and have the potential to 65 travel over continental or transoceanic scales fuelled by trade winds or dust storms (Hua et al., 66 2007; Prospero et al., 2005). During the airborne journey, bacteria are typically co-transported 67 embedded within organic detritus and/or soil particles that provide UV shielding and a certain 68 level of humidity that may increase their viability and dispersal ranges (Griffin et al., 2017). 69 Hence, aerosol transport may represent a successful mechanism for worldwide dispersal of 70 viable microorganisms and genes (Hervàs et al., 2009; Zhu et al., 2017). Airborne particles may

2 ACCEPTED MANUSCRIPT 71 therefore act as very effective vectors for intercontinental disease transmission and spreading of 72 allergens. Moreover, their influence in atmospheric and climate processes, as condensation 73 nuclei for cloud droplets, ice crystals, and precipitation is well recognized (Delort et al., 2010; 74 Fröhlich-Nowoisky et al., 2016; Joly et al., 2013). 75 76 Sampling bioaerosols from the free troposphere has been historically challenging, and there is a 77 need of systematic bioaerosols studies across time and space in the high atmosphere (Smith, 78 2013a, Caliz et al. 2018) like it is usually carried out with other common air pollutants (Bianchi 79 et al., 2016; Murray et al., 2009). A number of potential options for large spatio-temporal 80 studies of free troposphere bioaerosols has been recently reviewed, including aircrafts sampling, 81 scientific balloons, rockets, and remote sensing satellites (Smith, 2013b). These methodologies 82 are, however, expensive, need of complex logistics, and usually yield limited microbial biomass 83 that require of extensive controls and careful sterilization procedures. Conversely, high altitude 84 research stations have been used to assess tropospheric background concentrations and long- 85 range transport of pollutants for decades (for instance, the Swiss station at Jungfraujoch and 86 Mount Washington in New Hampshire, USA) (Bianchi et al., 2016; Murray et al., 2009). 87 Sampling bioaerosols at high-elevation mountains (Barberán et al., 2014; Bowers et al., 2012, 88 2009; Caliz et al., 2018) and studying the composition of their natural atmospheric depositions 89 (Amato et al., 2017; Hervàs et al., 2009; Peter et al., 2014) may also offer realistic possibilities 90 for temporal dynamics and long-range studies of bio MANUSCRIPTaerosols mimicking the medium-to-upper 91 troposphere microbiome (DeLeon-Rodriguez et al., 2013). However, the vertical distribution of 92 airborne microbial communities in different atmospheric layers have been only locally studied 93 (Maki et al., 2015), and there is a lack of studies to determine whether or not collecting wet and 94 dry atmospheric depositions could be representative of the free troposphere bioaerosols. 95 96 In the present study, we analysed the bacterial bioaerosols deposited under wet (rain) and dry 97 conditions, which were passively collected at a high-elevation mountain (Sierra Nevada, SE 98 Spain, 2900 m above sea level) along a set of sampling consecutive periods (over summer to 99 mid-autumn) lasting 3 years by high-throughput sequencing of 16S rRNA gene amplicons. 100 Next, we carried out a meta-analysis combining the Sierra Nevada dataset with data from 101 different studiesACCEPTED in the literature, covering a wide range of environmental conditions, elevations, 102 and bioaerosol origins. We hypothesized a close relationship between bacterial bioaerosols 103 collected on the top of high mountains and the free troposphere bioaerosols. We compared data 104 from bacterial bioaerosols obtained in variety of situations: (i) above the boundary layer by 105 aircraft sampling (DeLeon-Rodriguez et al., 2013; Zweifel et al., 2012); and under the influence 106 of dust episodes (Maki et al., 2017; Yamaguchi et al., 2012); (ii) from the ground surface at 107 high altitude research stations by means of air dynamic samplers (Bowers et al., 2012), and

3 ACCEPTED MANUSCRIPT 108 passively collecting wet (Barberán et al., 2014) or bulk (Peter et al., 2014) depositions; (iii) 109 from the ground surface at low altitude by means of air dynamic samplers (Bertolini et al., 110 2013; Bowers et al., 2013; Jeon et al., 2011; Mazar et al., 2016; Rosselli et al., 2015) or 111 collecting wet depositions (Itani and Smith, 2016); and (iv) close to the ocean surface by means 112 of air dynamic samplers (Mayol et al., 2017). 113 114 115 2. Materials and Methods 116 2.1. Sampling settings 117 Atmospheric depositions were collected above tree line in Sierra Nevada (Southeast Spain, 118 37º03’N, 3º23’W) far away from vegetation, trees, or agricultural activities, using a passive 119 MTXH ARS 1010 automatic sampler (MTX, Bologna, Italy) installed at 2896 m above sea level 120 on a metallic structure with 1.1 m legs on rocky soil at the Astrophysical Observatory of Sierra 121 Nevada (See Figure S1 —supplementary information) surrounded by rocks and meadows. This 122 passive collector discriminated between dry and wet atmospheric deposition using a humidity 123 sensor that activated an aluminium lid to cover/uncover two containers (dry and wet, 124 respectively). In this area, the annual mean for the boundary layer was located at 1.7 ± 0.5 km 125 altitude during the studied period —as determined by LIDAR data— (Granados-Muñoz et al., 126 2012), showing that the samples were predominantlyMANUSCRIPT collected above the boundary layer. 127 Moreover, temporal evolution of LIDAR signals corre sponding to isolated atmospheric events 128 previously analysed in this area (Mladenov et al., 2010) supported this fact. This region is under 129 the influence of the global dust belt, and has frequent intrusions of Saharan dust (Mladenov et 130 al., 2011). The origin of air masses reaching the Sierra Nevada Mountains was determined using 131 the transport and dispersion Hybrid Single-Particle Lagrangian Integrated (HYSPLIT) model 132 (Draxler and Rolph, 2014). HYSPLIT backward trajectories were obtained using archived data 133 from the Global Data Assimilation System with a 120 h run time at 2896 m asl. In addition, to 134 verify the origin of the air masses, we also generated maps of time-averaged dust-column mass 135 density (kg m −2) for the study periods using the second Modern-Era Retrospective analysis for 136 Research and Applications (MERRA-2) model. More details on these two procedures can be 137 found in RecheACCEPTED et al. (2018). 138 139 Dry and wet (rain) deposition samples were collected along three summers to mid- 140 autumn periods, fortnightly during 2006 and 2007, and weekly during 2008 (Reche at al. 2018). 141 The sample set consisted of n=46 samples (28 from the dry container and 18 from the wet). Dry 142 deposition was obtained rinsing the dry bucket with 1000 mL of Milli-Q ultrapure, 0.2 µm- 143 filtered, and UV-sterilized water. In the wet deposition bucket, the volume of rain was recorded

4 ACCEPTED MANUSCRIPT 144 and up to 1000 mL aliquot was processed when available. Blank controls in the used Milli-Q 145 ultrapure water were below detection level after bacterial counts with flow cytometry (Reche et 146 al. 2018). For particulate material determinations, water samples with suspended materials were 147 filtered using a 0.2 µm policarbonate filter (47 mm diameter). Filtration was stopped when 0.2 148 µm filter was saturated. Filters containing microbial biomass were stored at -20ºC before DNA 149 extraction. The total iron content determinations in the collected aerosols were carried out 150 following standard protocols as described earlier (Mladenov et al., 2010). 151 152 153 2.2. DNA extraction and sequencing 154 DNA was extracted using the UltraClean Plant DNA Kit (Mobio Laboratories, Inc.) 155 according to manufacturer’s instructions. DNA concentrations were measured using QubitTM 156 fluorometer (Invitrogen). The purified DNA extracts were stored at -80°C until use. Bacterial 157 bioaerosols were analysed by PCR amplified 16S rRNA gene tag sequencing with the primer 158 pair 357F/926R (357F-CCTACGGGAGGCAGCAG, 926R-CCGTCAATTCMTTTRAGT) 159 matching the V3–V5 hypervariable regions (Sim et al., 2012) by 454 FLX following the 160 Research and Testing Laboratory protocols (Lubbock, TX, http://www.researchandtesting.com). 161 Quality checking and reads denoising were processed using the UPARSE pipeline (Edgar, 162 2013), version usearch8.1.1861_i86osx32. Sequences were trimmed in length, 250 pb, which 163 substantially reduce the error rate. Approximately 99,2%MANUSCRIPT of the original reads passed the filter 164 (145627 final reads). The sequences were then clustered into operational taxonomic units 165 (OTUs) using a cut-off of 0.03% and, further analysed with UCHIME to discard chimeric reads 166 by both de novo and reference-based chimera filtering step against “Gold” reference database 167 available for ChimeraSlayer. Unique sequences (i.e. singletons) were removed. Overall, 132957 168 sequencing reads were assigned to OTUs, which were then parsed to create an OTU table. An 169 average sequence depth of 2461 sequences per sample was reached. In order to minimize biased 170 effects for differences in sampling effort, the original OTU table was average rarefied after 100 171 random subsamplings (Caliz et al., 2015) and set to a depth of 900 sequences per sample. 172 Taxonomic assignment was carried out with SINA – Aligner and classifier (version 1.2.11) 173 using SILVA_128 reference database (Pruesse et al., 2012; Quast et al., 2013) with the de- 174 replicated versionACCEPTED at 99% sequence identity. Chloroplasts, mitochondria, and a few untargeted 175 species were also discarded from the analysis. As a result, the final OTU table contained 1469 176 bacterial OTUs, and 46 samples. The whole gene sequence dataset was uploaded to the NCBI 177 Sequence Read Archive facility and is available through BioProject record ID PRJNA509787. 178 Metadata summarizing air masses origins —based on the analyses of backward trajectories, as 179 reported by (Reche et al., 2018)— are available in the BioProject record. 180

5 ACCEPTED MANUSCRIPT 181 182 2.3. Statistical analyses 183 Statistical analyses were run in the R environment (http://www.r-project.org/). Community 184 ecology related parameters were calculated using the vegan package (Oksanen et al., 2017) and 185 figures were drawn with ggplot2 (Wickham, 2009). Community similarities were represented by 186 non-metric multidimensional scaling ( metaMDS function) using Bray-Curtis dissimilarities after 187 Hellinger standardization (Legendre and Gallagher, 2001). Analyses of similarities (ANOSIM) 188 were performed based on 1000 permutations. The ANOSIM R statistic is based on the 189 difference of mean ranks between groups and within groups and ranges from 0 (no separation) 190 to 1 (complete separation) (Clarke and Warwick, 1994). With the aim to rule out 191 heteroscedasticity among groups the permutest.betadisper function was check before proceeding 192 with ANOSIM. This measure was also useful to assess the beta-diversity. We used Mantel tests 193 with the Spearman method to determine the correlation between bacterial community similarity 194 patterns and numeric variables. Shannon index and expected species richness were calculated 195 with functions diversity and rarefy (vegan package), respectively. Similarly, Berger-Parker 196 Dominance index was calculated by using the diverse package (Guevara et al., 2016). Venn 197 diagrams were carried out with VennDiagram package (Chen, 2016). Ward's hierarchical 198 agglomerative method based on Bray-Curtis distances was used for cluster analysis (hclust 199 function). 200 MANUSCRIPT 201 2.4. Meta-analysis 202 We carried out a meta-analysis combining published data from different genetic studies of 203 bacterial bioaerosols (see Table 1 for details) including data from the Pyrenees and the Sierra 204 Nevada dataset. Taxonomic compositions of bioaerosols were compared for a range of sampling 205 sites, methods, elevation, origins and occurrence of dust events. For each bioaerosol study, we 206 computed a summarized taxonomic profile. We used hierarchical clustering analysis to group all 207 bacterial bioaerosol studies according to taxonomic similarity. 208 209 210 3. Results 211 Bacterial bioaerosolsACCEPTED from dry and wet depositions collected in Sierra Nevada were barely 212 separated as indicated by a rather low, though significant, ANOSIM statistic (R: 0.17; p < 0.01). 213 A high degree of community overlapping is assumed for R < 0.25 (Ramette, 2007). The NMDS 214 ordination analysis based on Bray-Curtis dissimilarities showed high similarity as well (Figure 1 215 – left panel). No differences in beta diversity were observed between dry and wet depositions 216 (Betadisper test, p-value > 0.05) (Figure 1– right panel). In addition, the amount of deposited

6 ACCEPTED MANUSCRIPT 217 particulate materials and Fe —a marker signature indicating desert soil origin— did not support 218 samples dissimilarities (Mantel r – Spearman: 0.13; Significance: 0.07) (Figure S2). This result 219 could be explained by the mostly mixed origin of the aerosols samples in our dataset. Both wet 220 and dry bioaerosols showed high values of Good’s coverage of the bacterial communities (> 221 0.9) (Figure 2), and no significant differences were observed in Shannon index and Berger- 222 Parker Dominance (T-test, p-value > 0.05). Species richness and Faith’s phylogenetic diversity 223 were significantly higher in dry bioaerosols (Figure 2), even though the degree of taxa 224 overlapping was substantially high between dry and wet depositions (Figure S3). Up to 65 % of 225 OTUs and 82 % of the bacterial genera were shared between wet and dry bioaerosols, and the 226 dominant general taxa (i.e., relative abundances > 1%) were mainly the same (Figure S4, upper 227 panel) and positively correlated (Spearman rho: 0.54; p-value < 0.01) at the family level. 228 Overall, a combined dispersion plot for the mean relative abundances showed largely 229 convergent bioaerosols compositions at the taxonomic level of family (Figure S4, lower panel). 230 Interestingly, only Oxalobacteraceae were notably more abundant in wet deposition, with 231 Noviherbaspirillum and Massilia as dominant genera. 232 233 We carried out a meta-analysis according to taxonomic similarities combining the Sierra 234 Nevada dataset with data from different bioaerosols studies in a variety of situations and 235 sampling protocols including those from the free troposphere and the Pyrenees (Table 1). 236 Hierarchical clustering analysis divided the datase t MANUSCRIPTinto three main groups, named 1, 2a and 2b 237 in Figure 3. Cluster 1 included most of the studies within the boundary layer from low-elevation 238 areas and desert dust plumes collected from the free troposphere by aircrafts, owning a 239 substantial influence of ground materials. These samples showed a strong dominance of 240 Firmicutes sequences (relative abundance 31.6 ± 13.0 in cluster 1 vs 3.6 ± 3.5 and 9.4 ± 5.9 in 241 clusters 2a and 2b, respectively). Conversely, cluster 2a included bioaerosols mostly collected 242 above the boundary layer, both by aircrafts during desert dust free episodes and by passive 243 deposition at high-elevated mountains. Cluster 2b contained a heterogeneous variety of 244 sampling protocols and presence of dust, mainly representing local terrestrial inputs sampled by 245 air dynamic methods both at high altitudes and in suburban localities, as well as marine 246 aerosols. Bacterial communities from clusters 2a and 2b were more similar among them, with 247 higher proportionACCEPTED of Proteobacteria sequences than cluster 1 and always dominated by α-, β- 248 and γ-Proteobacteria. α-Proteobacteria clearly dominated cluster-2a (45.8 ± 11.3), while cluster 249 2b showed a higher mean relative abundance of β- (32.0 ± 22.0) and γ-Proteobacteria (17.7 ± 250 13.8). 251 252 The different studies grouped in cluster 2a were also highly coincident in bacterial composition 253 and structure at the family taxonomic level (Figure 4). The most abundant families such as

7 ACCEPTED MANUSCRIPT 254 Sphingomonadaceae / Methylobacteriaceae ( α-Proteobacteria) and Oxalobacteraceae / 255 Comamonadaceae ( β-Proteobacteria) were recurrently detected in all studies. 256 Pseudomonadaceae ( γ-Proteobacteria) were also abundant in several studies and almost evenly 257 detected. ANOSIM statistic further confirmed the high degree of overlapping in bacterial 258 composition between aircrafts sampling and passive deposition at high-elevated mountains (R: 259 0.15; significance: 0.2). 260 261 262 263 4. Discussion 264 Tropospheric communities are expected to be less complex than many of the biological habitats 265 on Earth (DeLeon-Rodriguez et al., 2013), and the meta-analysis of microbial communities 266 from a communal Earth' samples catalogue (Thompson et al., 2017) has shown lower alpha- 267 diversity in bioaerosols than in soils, sediments, plants rhizosphere and freshwaters. Wet 268 depositions collected at a high-elevated mountain in the Pyrenees also showed not very diverse 269 bacterial communities in comparison with source Saharan desert soils, and alpine lakes neuston 270 at the air-water interface (Barberán et al., 2014). In our study, the ecological diversity and 271 specific richness were quite low for most of the collected bacterial bioaerosols. Shannon index 272 (H’) values < 2 for all the cases indicate both low richnessMANUSCRIPT and high dominance of a small 273 number of taxa. This fact contrasts with the high d iversity of bioaerosols collected by air 274 dynamics near the ground (Bowers et al., 2012; Brodie et al., 2007) where most probably 275 contamination of the bioaerosol fraction with local ground surface microbes increases diversity. 276 Indeed, it has been shown that bacterial bioaerosols diversity increases due to the presence of 277 desert dust inputs both in the free troposphere and near to the ground (Maki et al., 2017; Mazar 278 et al., 2016). 279 280 As a relevant finding, we demonstrated that the bacterial composition of bioaerosols collected 281 by passive natural deposition at high-elevated mountains were closer to bacterial communities 282 from the free troposphere. Consequently, sampling alpine bioaerosols could be a good proxy for 283 bioaerosols monitoring and long-range dispersal studies. In fact, it has been recently shown a 284 closer relationshipACCEPTED between bioaerosols collected by air sampling at the top of mountains and at 285 the medium-to-upper troposphere than with those collected closer to oceanic or soil sources 286 (DeLeon-Rodriguez et al., 2013). As shown in our meta-analysis, these alpine airborne 287 communities share major orders, families and genera with the free troposphere. More in detail, 288 shared taxa were Sphingomonadaceae ( Sphingomonas ), Methylobacteriaceae 289 (Methylobacterium ), Oxalobacteraceae ( Massilia , Herbaspirillum , Noviherbaspirillum ),

8 ACCEPTED MANUSCRIPT 290 Pseudomonadaceae ( Pseudomonas ) and also other Rhizobiales. Besides, some of these taxa 291 include members related to known bacteria able to use C1-C4 compounds or to resist UV and 292 desiccation, which have been recurrently detected in a long-term monitoring of aerosols in a 293 mountain-top study in the Pyrenees (Caliz et al., 2018), evidencing their ability to potentially 294 survive and remain alive in the atmosphere. Then, our data is comparable to the results obtained 295 by means of sampling on an aircraft with a dynamic air setting (DeLeon-Rodriguez et al., 2013; 296 Maki et al., 2017; Zweifel et al., 2012), and the homogeneity of airborne communities could be 297 attributed, in part, to the different biological response of bacteria to cope with the atmospheric 298 environmental conditions. It is an obvious concern that research aiming to address long-range 299 airborne microbial dispersal should carefully avoid ground-surface contamination, and collect 300 samples from the free troposphere (Maki et al., 2017; Smith, 2013b). Since the boundary layer 301 can be easily reached in high mountain areas, and the local landscape is surrounded by rocks 302 and meadows, alpine stations are potentially optimal research sites with reduced influence of 303 surface aerosols, minimizing local contaminations. Subsequently, high mountain sites appear as 304 suitable sampling network infrastructures to address studies on the composition and fate of 305 long-range transported bioaerosols. The links of bacterial origins with trajectories is certainly an 306 issue of major interest that has been recently addressed in a high mountain station in the 307 Pyrenees, using a large number of samples within a 7 yrs. long-term study (Caliz et al. 2018) 308 309 Interestingly, the meta-analysis showed a different bacterialMANUSCRIPT composition and community 310 structure in bacterial bioaerosols collected at low-elevated areas, or during desert dust events, 311 with dominance of Firmicutes, Actinobacteria and Bacteroidetes. These phyla are commonly 312 found in soils (Fierer, 2017), and prevailed in both the bioaerosols associated with Asian dust 313 events and soils from different Asian deserts (Jeon et al., 2011; Maki et al., 2017). These 314 observations indicate that ground surface microbes can greatly influence the bioaerosols 315 collected at low-elevated sites, or during dust intrusion episodes (see listed works in cluster 1, 316 Fig. 3). In agreement, bioaerosols collected over the open ocean showed lower proportion of 317 these phyla and greater signal of terrestrial microbes are detected in marine bioaerosols 318 collected closer to landmasses than in the open ocean (Mayol et al., 2017). 319 320 In addition, samplingACCEPTED of passively deposited aerosols results advantageous in comparison with 321 air dynamic sampling on ground, as it would collect the microbiome present in the entire air 322 column above the boundary layer. Yet, we cannot obviate some potential biases and 323 contaminations. For instance, ground surface microbes deposited into the collectors from 324 surrounding areas by turbulent flows could represent a source of contamination. However, the 325 location of the collector above tree line and on rocky naked soil, and far away from vegetation, 326 trees, or agricultural activities strongly minimized this potential source of contamination. MilliQ

9 ACCEPTED MANUSCRIPT 327 water blanks were used to control for potential contaminations in dry depositions, that was 328 negligible. For wet deposition, we did not expect substantial enrichment of bacteria inside the 329 collector according to previous tests and test carried out in samples of the same dataset (Reche 330 et al., 2018). Thus, wet and the dry deposition rates for bacteria were not significantly different 331 (Kruskal–Wallis test KW-H = 2.88; p = 0.089, Reche et al. 2018), supporting no substantial 332 growth in the wet collector. Nevertheless, the use of bacterial growth inhibitors could be an 333 appropriate strategy to minimize potential biases during monitoring and handling of bioaerosols, 334 together with the synchrony in atmospheric deposition among distant sites (Reche et al., 2018). 335 Passive sampling of natural depositions allows to minimize maintenance logistics and powering 336 of sampling instrumentation, as there is no need for pumps in continuous running —either for 337 filtration settings or liquid impingers. 338 339 340 341 5. Conclusions 342 Overall, we observed that naturally deposited aerosols in areas with little influence of ground 343 surface contamination, such as the top of high-mountains, strongly resembled the bioaerosols of 344 the free troposphere. Thereby, this approach would be useful on reaching a descriptive view of 345 microbes that move thousands of kilometres away froMANUSCRIPTm source regions throughout the 346 troposphere. Under certain circumstances, wet collectors can recover higher bacterial loads as 347 compared with dry ones, for instance, during Saharan dust intrusions (Reche et al., 2018). Then, 348 long term monitoring studies employing molecular methods could be more easily affordable by 349 means of collecting bioaerosols washed from the atmosphere. This would allow circumventing 350 the problem of limited biomass, which could hamper bacterial aerosols studies. The proposal of 351 monitoring aeroplankton is noteworthy at the light of statements in need of monitoring microbes 352 just like usually done with other types of common air pollutants (Smith, 2013a, Caliz et al. 353 2018). Then, cost-effective, and long-term sustainable monitoring programs are needed for this 354 aim. Monitoring naturally deposited bioaerosols at accessible high-elevated areas would be a 355 good enough method for a realistic and general prospection of the free troposphere microbiome. 356 ACCEPTED 357 Acknowledgements 358 We are thankful to the authorities of the Sierra Nevada National Park for sampling facilities in 359 the protected areas, to Jorge Ramos and Natalie Mladenov for assistance with samples 360 collection, and to projects ECOSENSOR-BIOCON 04/009 from Fundación BBVA and

10 ACCEPTED MANUSCRIPT 361 DISPERSAL 829/2013 from the Spanish Office for the Environment-National Parks Research 362 Network (MAPAMA-OAPN to EOC). 363 364 References 365 Amato, P., Brisebois, E., Draghi, M., Duchaine, C., Fröhlich-Nowoisky, J., Huffman, 366 J.A., Mainelis, G., Robine, E., Thibaudon, M., 2017. Sampling Techniques, in: 367 Microbiology of Aerosols. Wiley-Blackwell, pp. 23–48. 368 https://doi.org/10.1002/9781119132318.ch1b 369 Barberán, A., Henley, J., Fierer, N., Casamayor, E.O., 2014. Structure, inter-annual 370 recurrence, and global-scale connectivity of airborne microbial communities. Sci. 371 Total Environ. 487, 187–195. 372 https://doi.org/https://doi.org/10.1016/j.scitotenv.2014.04.030 373 Bertolini, V., Gandolfi, I., Ambrosini, R., Bestetti, G., Innocente, E., Rampazzo, G., 374 Franzetti, A., 2013. Temporal variability and effect of environmental variables on 375 airborne bacterial communities in an urban area of Northern Italy. Appl. Microbiol. 376 Biotechnol. 97, 6561–6570. https://doi.org/10.1007/s00253-012-4450-0 377 Bianchi, F., Tröstl, J., Junninen, H., Frege, C., Henne, S., Hoyle, C.R., Molteni, U., 378 Herrmann, E., Adamov, A., Bukowiecki, N., Chen, X., Duplissy, J., Gysel, M., 379 Hutterli, M., Kangasluoma, J., Kontkanen, J., Kürten, A., Manninen, H.E., Münch, 380 S., Peräkylä, O., Petäjä, T., Rondo, L., Williamson, C., Weingartner, E., Curtius, J., 381 Worsnop, D.R., Kulmala, M., Dommen, J., Baltensperger, U., 2016. New particle 382 formation in the free troposphere: A question of chemistry and timing. Science 383 (80-. ). https://doi.org/10.1126/science.aad5456 384 Bowers, R.M., Clements, N., Emerson, J.B., Wiedinmyer, C., Hannigan, M.P., Fierer, 385 N., 2013. Seasonal Variability in BacterialMANUSCRIPT and Fungal Diversity of the Near- 386 Surface Atmosphere. Environ. Sci. Technol. 47, 12097–12106. 387 https://doi.org/10.1021/es402970s 388 Bowers, R.M., Lauber, C.L., Wiedinmyer, C., Hamady, M., Hallar, A.G., Fall, R., 389 Knight, R., Fierer, N., 2009. Characterization of airborne microbial communities at 390 a high-elevation site and their potential to act as atmospheric ice nuclei. Appl. 391 Environ. Microbiol. 75, 5121–5130. https://doi.org/10.1128/AEM.00447-09 392 Bowers, R.M., McCubbin, I.B., Hallar, A.G., Fierer, N., 2012. Seasonal variability in 393 airborne bacterial communities at a high-elevation site. Atmos. Environ. 50, 41–49. 394 https://doi.org/10.1016/j.atmosenv.2012.01.005 395 Brodie, E.L., DeSantis, T.Z., Parker, J.P.M., Zubietta, I.X., Piceno, Y.M., Andersen, 396 G.L., 2007. Urban aerosols harbor diverse and dynamic bacterial populations. 397 Proc. Natl. Acad. Sci. 104, 299–304. https://doi.org/10.1073/pnas.0608255104 398 Burrows, S.M., Elbert, W., Lawrence, M.G., Pöschl, U., 2009. Bacteria in the global 399 atmosphere - Part 1: Review and synthesis of literature data for different 400 ecosystems.ACCEPTED Atmos. Chem. Phys. 9, 9263–9280. https://doi.org/10.5194/acp-9- 401 9263-2009 402 Caliz, J., Montes-Borrego, M., Triadó-Margarit, X., Metsis, M., Landa, B.B., 403 Casamayor, E.O., 2015. Influence of Edaphic, Climatic, and Agronomic Factors on 404 the Composition and Abundance of Nitrifying Microorganisms in the Rhizosphere 405 of Commercial Olive Crops. PLoS One 10, 1–17. 406 https://doi.org/10.1371/journal.pone.0125787 407 Caliz, J., Triadó-Margarit, X., Camarero, L., Casamayor, E.O., 2018. A long-term 408 survey unveils consistent seasonal trends in the airborne microbiome composition

11 ACCEPTED MANUSCRIPT 409 coupled to air masses circulation. Proc. Natl. Acad. Sci. in press. 410 https://doi.org/10.1073/pnas.1812826115 411 Chen, H., 2016. VennDiagram: Generate High-Resolution Venn and Euler Plots. 412 Clarke, K.R., Warwick, R.M., 1994. Similarity-based testing for community pattern: the 413 two-way layout with no replication. Mar. Biol. 118, 167–176. 414 https://doi.org/10.1007/BF00699231 415 DeLeon-Rodriguez, N., Lathem, T.L., Rodriguez-R, L.M., Barazesh, J.M., Anderson, 416 B.E., Beyersdorf, A.J., Ziemba, L.D., Bergin, M., Nenes, A., Konstantinidis, K.T., 417 2013. Microbiome of the upper troposphere: Species composition and prevalence, 418 effects of tropical storms, and atmospheric implications. Proc. Natl. Acad. Sci. 419 110, 2575–2580. https://doi.org/10.1073/pnas.1212089110 420 Delort, A.-M., Vaïtilingom, M., Amato, P., Sancelme, M., Parazols, M., Mailhot, G., 421 Laj, P., Deguillaume, L., 2010. A short overview of the microbial population in 422 clouds: Potential roles in atmospheric chemistry and nucleation processes. Atmos. 423 Res. 98, 249–260. https://doi.org/https://doi.org/10.1016/j.atmosres.2010.07.004 424 Edgar, R.C., 2013. UPARSE: highly accurate OTU sequences from microbial amplicon 425 reads. Nat Methods 10, 996–998. https://doi.org/10.1038/nmeth.2604 426 Fierer, N., 2017. Embracing the unknown: Disentangling the complexities of the soil 427 microbiome. Nat. Rev. Microbiol. 15, 579–590. 428 https://doi.org/10.1038/nrmicro.2017.87 429 Fröhlich-Nowoisky, J., Kampf, C.J., Weber, B., Huffman, J.A., Pöhlker, C., Andreae, 430 M.O., Lang-Yona, N., Burrows, S.M., Gunthe, S.S., Elbert, W., Su, H., Hoor, P., 431 Thines, E., Hoffmann, T., Després, V.R., Pöschl, U., 2016. Bioaerosols in the 432 Earth system: Climate, health, and ecosystem interactions. Atmos. Res. 182, 346– 433 376. https://doi.org/10.1016/j.atmosres.2016.07.018 434 Granados-Muñoz, M.J., Navas-Guzmán, F., Bravo-Aranda, J.A., Guerrero-Rascado, 435 J.L., Lyamani, H., Fernández-Gálvez, J., Alados-ArbMANUSCRIPToledas, L., 2012. Automatic 436 determination of the planetary boundary layer height using lidar: One-year analysis 437 over southeastern Spain. J. Geophys. Res. Atmos. 117. 438 https://doi.org/10.1029/2012JD017524 439 Griffin, D.W., Gonzalez-Martin, C., Hoose, C., Smith, D.J., 2017. Global-Scale 440 Atmospheric Dispersion of Microorganisms, in: Microbiology of Aerosols. Wiley- 441 Blackwell, pp. 155–194. https://doi.org/10.1002/9781119132318.ch2c 442 Guevara, M.R., Hartmann, D., Mendoza, M., 2016. diverse: an {R} {Package} to 443 {Measure} {Diversity} in {Complex} {Systems}. R J. 8, 60–78. 444 Hervàs, A., Camarero, L., Reche, I., Casamayor, E.O., 2009. Viability and potential for 445 immigration of airborne bacteria from Africa that reach high mountain lakes in 446 Europe. Environ. Microbiol. 11, 1612–1623. https://doi.org/10.1111/j.1462- 447 2920.2009.01926.x 448 Hua, N.-P., Kobayashi, F., Iwasaka, Y., Shi, G.-Y., Naganuma, T., 2007. Detailed 449 identification of desert-originated bacteria carried by Asian dust storms to Japan. 450 AerobiologiaACCEPTED (Bologna). 23, 291–298. https://doi.org/10.1007/s10453-007-9076-9 451 Itani, G.N., Smith, C.A., 2016. Dust Rains Deliver Diverse Assemblages of 452 Microorganisms to the Eastern Mediterranean. Sci. Rep. 6, 1–11. 453 https://doi.org/10.1038/srep22657 454 Jeon, E.M., Kim, H.J., Jung, K., Kim, J.H., Kim, M.Y., Kim, Y.P., Ka, J.O., 2011. 455 Impact of Asian dust events on airborne bacterial community assessed by 456 molecular analyses. Atmos. Environ. 45, 4313–4321. 457 https://doi.org/10.1016/j.atmosenv.2010.11.054 458 Joly, M., Attard, E., Sancelme, M., Deguillaume, L., Guilbaud, C., Morris, C.E., Amato,

12 ACCEPTED MANUSCRIPT 459 P., Delort, A.-M., 2013. Ice nucleation activity of bacteria isolated from cloud 460 water. Atmos. Environ. 70, 392–400. 461 https://doi.org/https://doi.org/10.1016/j.atmosenv.2013.01.027 462 Legendre, P., Gallagher, E.D., 2001. Ecologically meaningful transformations for 463 ordination of species data. Oecologia 129, 271–280. 464 https://doi.org/10.1007/s004420100716 465 Maki, T., Hara, K., Iwata, A., Lee, K.C., Kawai, K., Kai, K., Kobayashi, F., Pointing, 466 S.B., Archer, S., Hasegawa, H., Iwasaka, Y., 2017. Variations in airborne bacterial 467 communities at high altitudes over the Noto Peninsula (Japan) in response to Asian 468 dust events. Atmos. Chem. Phys. 17, 11877–11897. https://doi.org/10.5194/acp- 469 17-11877-2017 470 Maki, T., Hara, K., Kobayashi, F., Kurosaki, Y., Kakikawa, M., Matsuki, A., Chen, B., 471 Shi, G., Hasegawa, H., Iwasaka, Y., 2015. Vertical distribution of airborne 472 bacterial communities in an Asian-dust downwind area, Noto Peninsula. Atmos. 473 Environ. 119, 282–293. 474 https://doi.org/https://doi.org/10.1016/j.atmosenv.2015.08.052 475 Mayol, E., Arrieta, J.M., Jiménez, M.A., Martínez-Asensio, A., Garcias-Bonet, N., 476 Dachs, J., González-Gaya, B., Royer, S.J., Benítez-Barrios, V.M., Fraile-Nuez, E., 477 Duarte, C.M., 2017. Long-range transport of airborne microbes over the global 478 tropical and subtropical ocean. Nat. Commun. 8, 1–8. 479 https://doi.org/10.1038/s41467-017-00110-9 480 Mazar, Y., Cytryn, E., Erel, Y., Rudich, Y., 2016. Effect of Dust Storms on the 481 Atmospheric Microbiome in the Eastern Mediterranean. Environ. Sci. Technol. 50, 482 4194–4202. https://doi.org/10.1021/acs.est.5b06348 483 Mladenov, N., Reche, I., Olmo, F.J., Lyamani, H., Alados-Arboledas, L., 2010. 484 Relationships between spectroscopic properties of high-altitude organic aerosols 485 and Sun photometry from ground-based remoteMANUSCRIPT sensing. J. Geophys. Res. 486 Biogeosciences 115. https://doi.org/10.1029/2009JG0 00991 487 Mladenov, N., Sommaruga, R., Morales-Baquero, R., Laurion, I., Camarero, L., 488 Diéguez, M.C., Camacho, A., Delgado, A., Torres, O., Chen, Z., Felip, M., Reche, 489 I., 2011. Dust inputs and bacteria influence dissolved organic matter in clear alpine 490 lakes. Nat. Commun. 2, 405. 491 Murray, G.L.D., Kimball, K., Bruce Hill, L., Allen, G.A., Wolfson, J.M., Pszenny, A., 492 Seidel, T., Doddridge, B.G., Boris, A., 2009. A comparison of fine particle and 493 aerosol strong acidity at the interface zone (1540 m) and within (452 m) the 494 planetary boundary layer of the Great Gulf and Presidential-Dry River Class I 495 Wildernesses on the Presidential Range, New Hampshire USA. Atmos. Environ. 496 43, 3605–3613. https://doi.org/10.1016/j.atmosenv.2009.03.060 497 Oksanen, J., Blanchet, F.G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., 498 Minchin, P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H., 499 Szoecs, E., Wagner, H., 2017. vegan: Community Ecology Package. 500 Peter, H., Hörtnagl,ACCEPTED P., Reche, I., Sommaruga, R., 2014. Bacterial diversity and 501 composition during rain events with and without Saharan dust influence reaching a 502 high mountain lake in the Alps. Environ. Microbiol. Rep. 6, 618–624. 503 https://doi.org/10.1111/1758-2229.12175 504 Prospero, J.M., Blades, E., Mathison, G., Naidu, R., 2005. Interhemispheric transport of 505 viable fungi and bacteria from Africa to the Caribbean with soil dust. Aerobiologia 506 (Bologna). 21, 1–19. https://doi.org/10.1007/s10453-004-5872-7 507 Pruesse, E., Peplies, J., Glockner, F.O., 2012. SINA: accurate high-throughput multiple 508 sequence alignment of ribosomal RNA genes. Bioinformatics 28, 1823–1829.

13 ACCEPTED MANUSCRIPT 509 https://doi.org/10.1093/bioinformatics/bts252 510 Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., Peplies, J., 511 Glockner, F.O., 2013. The SILVA ribosomal RNA gene database project: 512 improved data processing and web-based tools. Nucleic Acids Res 41, D590-6. 513 https://doi.org/doi: 10.1093/nar/gks1219 514 Ramette, A., 2007. Multivariate analyses in microbial ecology. FEMS Microbiol. Ecol. 515 62, 142–160. https://doi.org/10.1111/j.1574-6941.2007.00375.x 516 Reche, I., D’Orta, G., Mladenov, N., Winget, D.M., Suttle, C.A., 2018. Deposition rates 517 of viruses and bacteria above the atmospheric boundary layer. The ISME 518 Journal 12 , 1154–1162 519 Rosselli, R., Fiamma, M., Deligios, M., Pintus, G., Pellizzaro, G., Canu, A., Duce, P., 520 Squartini, A., Muresu, R., Cappuccinelli, P., 2015. Microbial immigration across 521 the Mediterranean via airborne dust. Sci. Rep. 5, 1–10. 522 https://doi.org/10.1038/srep16306 523 Sim, K., Cox, M.J., Wopereis, H., Martin, R., Knol, J., Li, M.-S., Cookson, W.O.C.M., 524 Moffatt, M.F., Kroll, J.S., 2012. Improved Detection of Bifidobacteria with 525 Optimised 16S rRNA-Gene Based Pyrosequencing. PLoS One 7, 1–7. 526 https://doi.org/10.1371/journal.pone.0032543 527 Smith, D.J., 2013a. Aeroplankton and the Need for a Global Monitoring Network. 528 Bioscience 63, 515–516. https://doi.org/10.1525/bio.2013.63.7.3 529 Smith, D.J., 2013b. Microbes in the Upper Atmosphere and Unique Opportunities for 530 Astrobiology Research. Astrobiology 13, 981–990. 531 https://doi.org/10.1089/ast.2013.1074 532 Thompson, L.R., Sanders, J.G., McDonald, D., Amir, A., Ladau, J., Locey, K.J., Prill, 533 R.J., Tripathi, A., Gibbons, S.M., Ackermann, G., Navas-Molina, J.A., Janssen, S., 534 Kopylova, E., Vázquez-Baeza, Y., González, A., Morton, J.T., Mirarab, S., Xu, 535 Z.Z., Jiang, L., Haroon, M.F., Kanbar, J., Zhu,MANUSCRIPT Q., Song, S.J., Kosciolek, T., 536 Bokulich, N.A., Lefler, J., Brislawn, C.J., Humphrey, G., Owens, S.M., Hampton- 537 Marcell, J., Berg-Lyons, D., McKenzie, V., Fierer, N., Fuhrman, J.A., Clauset, A., 538 Stevens, R.L., Shade, A., Pollard, K.S., Goodwin, K.D., Jansson, J.K., Gilbert, 539 J.A., Knight, R., Agosto Rivera, J.L., Al-Moosawi, L., Alverdy, J., Amato, K.R., 540 Andras, J., Angenent, L.T., Antonopoulos, D.A., Apprill, A., Armitage, D., 541 Ballantine, K., Bárta, J., Baum, J.K., Berry, A., Bhatnagar, A., Bhatnagar, M., 542 Biddle, J.F., Bittner, L., Boldgiv, B., Bottos, E., Boyer, D.M., Braun, J., Brazelton, 543 W., Brearley, F.Q., Campbell, A.H., Caporaso, J.G., Cardona, C., Carroll, J.L., 544 Cary, S.C., Casper, B.B., Charles, T.C., Chu, H., Claar, D.C., Clark, R.G., Clayton, 545 J.B., Clemente, J.C., Cochran, A., Coleman, M.L., Collins, G., Colwell, R.R., 546 Contreras, M., Crary, B.B., Creer, S., Cristol, D.A., Crump, B.C., Cui, D., Daly, 547 S.E., Davalos, L., Dawson, R.D., Defazio, J., Delsuc, F., Dionisi, H.M., 548 Dominguez-Bello, M.G., Dowell, R., Dubinsky, E.A., Dunn, P.O., Ercolini, D., 549 Espinoza, R.E., Ezenwa, V., Fenner, N., Findlay, H.S., Fleming, I.D., Fogliano, V., 550 Forsman,ACCEPTED A., Freeman, C., Friedman, E.S., Galindo, G., Garcia, L., Garcia-Amado, 551 M.A., Garshelis, D., Gasser, R.B., Gerdts, G., Gibson, M.K., Gifford, I., Gill, R.T., 552 Giray, T., Gittel, A., Golyshin, P., Gong, D., Grossart, H.P., Guyton, K., Haig, S.J., 553 Hale, V., Hall, R.S., Hallam, S.J., Handley, K.M., Hasan, N.A., Haydon, S.R., 554 Hickman, J.E., Hidalgo, G., Hofmockel, K.S., Hooker, J., Hulth, S., Hultman, J., 555 Hyde, E., Ibáñez-Álamo, J.D., Jastrow, J.D., Jex, A.R., Johnson, L.S., Johnston, 556 E.R., Joseph, S., Jurburg, S.D., Jurelevicius, D., Karlsson, A., Karlsson, R., 557 Kauppinen, S., Kellogg, C.T.E., Kennedy, S.J., Kerkhof, L.J., King, G.M., Kling, 558 G.W., Koehler, A. V., Krezalek, M., Kueneman, J., Lamendella, R., Landon, E.M.,

14 ACCEPTED MANUSCRIPT 559 Lanede Graaf, K., LaRoche, J., Larsen, P., Laverock, B., Lax, S., Lentino, M., 560 Levin, I.I., Liancourt, P., Liang, W., Linz, A.M., Lipson, D.A., Liu, Y., Lladser, 561 M.E., Lozada, M., Spirito, C.M., MacCormack, W.P., MacRae-Crerar, A., Magris, 562 M., Martín-Platero, A.M., Martín-Vivaldi, M., Martínez, L.M., Martínez-Bueno, 563 M., Marzinelli, E.M., Mason, O.U., Mayer, G.D., McDevitt-Irwin, J.M., 564 McDonald, J.E., McGuire, K.L., McMahon, K.D., McMinds, R., Medina, M., 565 Mendelson, J.R., Metcalf, J.L., Meyer, F., Michelangeli, F., Miller, K., Mills, D.A., 566 Minich, J., Mocali, S., Moitinho-Silva, L., Moore, A., Morgan-Kiss, R.M., 567 Munroe, P., Myrold, D., Neufeld, J.D., Ni, Y., Nicol, G.W., Nielsen, S., Nissimov, 568 J.I., Niu, K., Nolan, M.J., Noyce, K., O’Brien, S.L., Okamoto, N., Orlando, L., 569 Castellano, Y.O., Osuolale, O., Oswald, W., Parnell, J., Peralta-Sánchez, J.M., 570 Petraitis, P., Pfister, C., Pilon-Smits, E., Piombino, P., Pointing, S.B., Pollock, F.J., 571 Potter, C., Prithiviraj, B., Quince, C., Rani, A., Ranjan, R., Rao, S., Rees, A.P., 572 Richardson, M., Riebesell, U., Robinson, C., Rockne, K.J., Rodriguezl, S.M., 573 Rohwer, F., Roundstone, W., Safran, R.J., Sangwan, N., Sanz, V., Schrenk, M., 574 Schrenzel, M.D., Scott, N.M., Seger, R.L., Seguinorlando, A., Seldin, L., Seyler, 575 L.M., Shakhsheer, B., Sheets, G.M., Shen, C., Shi, Y., Shin, H., Shogan, B.D., 576 Shutler, D., Siegel, J., Simmons, S., Sjöling, S., Smith, D.P., Soler, J.J., Sperling, 577 M., Steinberg, P.D., Stephens, B., Stevens, M.A., Taghavi, S., Tai, V., Tait, K., 578 Tan, C.L., Ta ş, N., Taylor, D.L., Thomas, T., Timling, I., Turner, B.L., Urich, T., 579 Ursell, L.K., Van Der Lelie, D., Van Treuren, W., Van Zwieten, L., Vargas- 580 Robles, D., Thurber, R.V., Vitaglione, P., Walker, D.A., Walters, W.A., Wang, S., 581 Wang, T., Weaver, T., Webster, N.S., Wehrle, B., Weisenhorn, P., Weiss, S., 582 Werner, J.J., West, K., Whitehead, A., Whitehead, S.R., Whittingham, L.A., 583 Willerslev, E., Williams, A.E., Wood, S.A., Woodhams, D.C., Yang, Y., Zaneveld, 584 J., Zarraonaindia, I., Zhang, Q., Zhao, H., 2017. A communal catalogue reveals 585 Earth’s multiscale microbial diversity. NatureMANUSCRIPT 551, 457–463. 586 https://doi.org/10.1038/nature24621 587 Wickham, H., 2009. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New 588 York. 589 Yamaguchi, N., Ichijo, T., Sakotani, A., Baba, T., Nasu, M., 2012. Global dispersion of 590 bacterial cells on Asian dust. Sci. Rep. 2, 1–6. https://doi.org/10.1038/srep00525 591 Zhu, Y.-G., Gillings, M., Simonet, P., Stekel, D., Banwart, S., Penuelas, J., 2017. 592 Microbial mass movements. Science (80-. ). 357, 1099–1100. 593 https://doi.org/10.1126/science.aao3007 594 Zweifel, U.L., Hagström, Å., Holmfeldt, K., Thyrhaug, R., Geels, C., Frohn, L.M., 595 Skjøth, C.A., Karlson, U.G., 2012. High bacterial 16S rRNA gene diversity above 596 the atmospheric boundary layer. Aerobiologia (Bologna). 28, 481–498. 597 https://doi.org/10.1007/s10453-012-9250-6 598 599 ACCEPTED

15 ACCEPTED MANUSCRIPT 600 601

MANUSCRIPT

ACCEPTED

16 ACCEPTED MANUSCRIPT 602 Figure legends 603 604 Figure 1 Non-metric multidimensional scaling (NMDS) ordination analysis of bacterial 605 community similarities based on Bray-Curtis dissimilarities (left panel). Dots colour according 606 to the type of deposition. Dotted lines join each sample with its corresponding group centroid. 607 The ellipsoid areas are indicative of distribution (and dispersion) of each group on the 608 ordination space (at 0.95 confidence level). The right panel shows boxplots of beta-diversity as 609 distance to centroid. 610 611 Figure 2 Box-plots comparing the values of good’s coverage and a set of different alpha- 612 diversity parameters for dry and wet deposition. 613 614 Figure 3. Hierarchical clustering analysis (Ward's hierarchical agglomerative clustering 615 method) based on Bray-Curtis distances of bacterial bioaerosols reported from different studies. 616 Bar plots show the relative abundance of dominant phyla or classes. Information on sampling 617 sites, sampling techniques, and presence of desert dust is shown. See additional details in Table 618 1. 619 620 Figure 4 . Composition and abundance of taxonomicMANUSCRIPT bacterial families in bioaerosols from 621 studies in cluster 2a of Figure 3. Only families with relative abundances > 1% are shown. Total 622 abundance percentage analysed shown within brackets in the colour legend. Labels for sampling 623 strategy (upper part): arrow, deposition collection on ground surface at high altitude sites; 624 airplane, aircrafts equipped with air dynamic samplers (see also Table 1). 625

ACCEPTED

17 ACCEPTED MANUSCRIPT Table 1. Studies included in the meta-analysis. Details on geographic and sampling sites, elevation, sampling technique and occurrence of dust events are shown. The right-most column shows position in the cluster analysis shown in Fig. 4. ABL above boundary layer, BL boundary layer, n /a non-applicable, NS non-specified. * sampling site ABL just during night-time. Geographic Atmospheric Sampling Dust Cluster Reference Sampler Deposition site layer point Episode Fig. 4 Sardinia-Italy Ground surface (Rosselli et al., 2015) BL Air dynamic n/a No 1 (EU) - low altitude (Yamaguchi et al., Air dynamic Japan (Asia) ABL Air n/a Yes 1 2012) - Aircraft Sardinia-Italy Ground surface (Rosselli et al., 2015) BL Air dynamic n/a Yes 1 (EU) - low altitude Ground surface (Jeon et al., 2011) Seoul (Asia) BL Air dynamic n/a Yes 1 - low altitude Air dynamic (Maki et al., 2017) Japan (Asia) ABL Air n/a Yes 1 - Aircraft Ground surface (Mazar et al., 2016) Israel (Asia) BL Air dynamic n/a No 1 - low altitude Colorado Ground surface (Bowers et al., 2013) BL Air dynamic n/a NS 1 (USA) - low altitude Ground surface (Mazar et al., 2016) Israel (Asia) BL Air dynamic n/a Yes 1 - low altitude (Bertolini et al., Ground surface Italy (EU) BL Air dynamic n/a NS 1 2013) - low altitude Ground surface This study Spain (EU) ABL Deposition Wet n/a 2a - high altitude MANUSCRIPT (Barberán et al., Ground surface Spain (EU) ABL Deposition Wet NS 2a 2014) - high altitude (DeLeon-Rodriguez Air dynamic USA ABL Air n/a NS 2a et al., 2013) - Aircraft Air dynamic (Zweifel et al., 2012) Sweden (EU) ABL Air n/a NS 2a - Aircraft Ground surface This study Spain (EU) ABL Deposition Dry n/a 2a - high altitude Air dynamic (Maki et al., 2017) Japan (Asia) ABL Air n/a No 2a - Aircraft (Itani and Smith, Ground surface Beirut (Asia) BL Deposition Wet Yes 2b 2016) - low altitude Ground surface (Peter et al., 2014) Austria (EU) nd Deposition Bulk No 2b - high altitude ACCEPTEDColorado Ground surface (Bowers et al., 2012) BL* Air dynamic n/a NS 2b (USA) - high altitude Ocean (Mayol et al., 2017) BL Ocean surface Air dynamic n/a NS 2b (Global) Ground surface (Peter et al., 2014) Austria (EU) nd Deposition Bulk Yes 2b - high altitude Ground surface (Jeon et al., 2011) Seoul (Asia) BL Air dynamic n/a No 2b - low altitude

ACCEPTED MANUSCRIPT

DepositionCollector Dry Wet t ypeType

0.4

0.7

0.2

0.6

0.0 MANUSCRIPT

Axis 2

0.5

−0.2 Distance to centroid

0.4

−0.4 ACCEPTED −0.4 −0.2 0.0 0.2 0.4 Axis 1 ANOSIM R:0.168 Betadisper test: Significance:0.002 NS, p−value >0.05 ACCEPTED MANUSCRIPT

300 1.0 5

ersity 20 0.9 0.75 4 ichness age

x − H' 200 e r 15 v 0.8 3 er Dominance

k 0.50

0.7 a r 2 ylogenetic Di v 100 10

Good's c o 0.25 MANUSCRIPT 0.6 Shannon ind e 1 5 Rarefied species r Berger− P aith's P h

0.5 F 0 DRY WET DRY WET DRY WET DRY WET DRY WET Deposition type T−test T−test T−test, T−test, (log transform.), (log transform.), p−value > 0.05ACCEPTEDp−value > 0.05 p−value < 0.05 p−value < 0.05 ACCEPTED MANUSCRIPT

2.0 1.5 1.0 0.5 0 0 25 50 75 100 BoundaryAboveAir layer bondary- GroundusingGround aircraft layersurfaceOcean surfaceAir surface-dynamicDeposition high -Wet low altitude Drydepositionaltitude depositionDustNo episode dust episode Jeon et al., 2011 Cluster-2b Peter et al., 2014 Mayol et al., 2017 Bowers et al., 2012 Peter et al., 2014

Itani and Smith 2016 Pylum / Class Maki et al., 2017 Acidobacteria

Cluster-2a Actinobacteria This study Bacteroidetes Zweifel et. al 2012 Chloroflexi MANUSCRIPTDeLeon-Rodriguez Firmicutes et al. 2013 Cyanobacteria Barberan et al., 2014 Deinococcus Thermus This study α - Proteobacteria Bertolini et al., 2013 β - Proteobacteria Mazar et al., 2016 δ - Proteobacteria ε - Proteobacteria Bowers et al., 2013 Cluster-1 γ - Proteobacteria Mazar et al., 2016 Others Maki et al., 2017 ACCEPTED Jeon et al., 2011 Rosselli et al., 2015 Yamaguchi et al., 2012 Rosselli et al., 2015 0 25 50 75 100 Sampling Sampling Relative abundance (%) site method ACCEPTED MANUSCRIPT

Tenericutes Spiroplasmataceae ● ● Yersiniaceae ● Gammaprot. Xanthomonadaceae ● ● ● ● Sinobacteraceae ● Shewanellaceae ● Pseudomonadaceae ● ● ● ● ● Legionellaceae ● ● Erwiniaceae ● ● ● ● Enterobacteriaceae Abundance (%) Deltaprot. Desulfobacteraceae ● Oxalobacteraceae ● ● ● ● ● ● ● 0.1 Betaprot. Comamonadaceae ● ● ● ● ● ● Burkholderiaceae ● ● ● ● ● ● 1 Sphingomonadaceae ● ● ● ● ● Alphaprot. ● ● Rhodospirillaceae ● ● ● 5 Rhodobacteraceae ● ● ● ● ● ● ● ● ● Rhizobiaceae ● 10 Phyllobacteriaceae ● ● 20 Methylobacteriaceae ● ● ● ● ● ● ● ● ● Hyphomicrobiaceae 40 Caulobacteraceae ● ● ● ● ● ● Bradyrhizobiaceae ● ● ● Acetobacteraceae ● ●●● ● ● ● ● Betaprot. Staphylococcaceae Paenibacillaceae MANUSCRIPT● ● ● A. This study (wet Lactobacillaceae ● ● deposition) (94.2 %) Clostridiaceae ● ● ● B. Barberan et al., 2014 Bacillaceae ● ● ● ● ● ● (85.8 %) ● ● ● ● Deinococcus-Thermus Deinococcaceae C. DeLeon-Rodriguez Synechococcaceae ● ● et al., 2013 (78.1 %) Cyanobacteria Cyanobacteriaceae ● ● ● ● ● D. Zweifel el al., 2012 Sphingobacteriaceae ● (72.8 %) Bacteroidetes Flexibacteraceae ● ● ● ● E. This study (dry Flavobacteriaceae deposition) (84.6 %) Cytophagaceae ● ● ● Chitinophagaceae ● ● ● F. Maki et al., 2017 (89.9 %) Armatimonadetes Aurantimonadaceae ● ● ● ● ● ● Actinobacteria Propionibacteriaceae Nocardioidaceae ● ● ● ● ● Micrococcaceae ● ● ● ● ● Microbacteriaceae ● ● ● ● ● ACCEPTEDGeodermatophilaceae ● ● ● ACK−M1 ● Acidobacteria Acidobacteriaceae ● ● A B C D E F ACCEPTED MANUSCRIPT Highlights - Above the boundary layer ground-sampled bacterial bioaerosols were largely convergent with those from the free troposphere, both in composition and structure

- Bioaerosols passively deposited at high-elevation mountains are a good proxy for long- term monitoring of the high atmosphere long-range transported microbiome

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

MANUSCRIPT

ACCEPTED