Gut microbiome diversity across seasons and locations in thai captive Asian elephants (elephas maximus)

The gut microbiome plays a critical role in nutrient metabolism, immune function, and overall health of animals, particularly in herbivores including hindgut fermenters^1,2,3. However, its dynamics in captive Asian elephants (Elephas maximus) remain limited. Recent advancements in sequencing technologies have significantly improved our ability to investigate host-microbe interactions; however, the spatiotemporal dynamics of the gut microbiota—key determinants of host health in captive management contexts—remain insufficiently characterized in this endangered species.

In captive Asian elephants, preliminary studies have identified age^4,5– and diet^6,7-driven shifts in gut microbial communities. However, these investigations have primarily focused on single time points or locations, leaving critical gaps in understanding how seasonal resource availability and regional environmental conditions shape microbial ecology. For herbivores, seasonal dietary shifts are known to induce microbial turnover, as demonstrated in forest musk deer (Moschus berezovskii), where dry-leaf winter diets were correlated with increased microbial diversity and Firmicutes–Bacteroidetes ratios compared to fresh-leaf summer diets⁸. Similarly, wild herbivores, for example, goitered gazelles (Gazella subgutturosa) exhibit adaptive microbial restructuring during winter, favoring taxa that increase energy utilization from scarce resources⁹. These patterns suggest that captive elephants-whose diets are partially controlled but still subject to seasonal forage variations-may experience analogous microbial adaptations. Geographic location further compounds these dynamics through influencing the types of plants and dietary resources available to animals¹⁰. Furthermore, environmental microbes from soil, water, and vegetation can act as sources of colonization for the animal gut microbiome¹¹. For Thailand’s captive elephant populations, which are distributed across ecologically distinct regions, management practices and localized soil conditions could drive divergent microbial profiles. Despite these implications, no prior work has simultaneously assessed seasonal and location effects on Asian elephant gut microbiomes-a knowledge gap critical for optimizing conservation strategies and captive welfare.

This study investigated gut microbiomes of captive Asian elephants from two provinces in Thailand, Lampang (northern region) and Kanchanaburi (western region), sampled during wet and dry seasons. Using 16 S rRNA gene sequencing, the study characterized microbial diversity, taxonomic composition, and predicted functions, testing the hypotheses that (1) location-specific management practices and environmental conditions generate distinct microbial signatures independent of season and (2) seasonal dietary shifts drive major reorganizations of the gut microbiota. By integrating spatial and temporal dimensions, these findings provide insights into microbial resilience in captive elephants and offer potential targets for management and dietary interventions aimed at improving health and welfare.

The composition of the gut microbiome

Sequencing of each sample yielded an average ± SD of 101,083.54 ± 620.11 reads, with the number of reads per sample ranging from 100,018 to 101,965. These reads were classified into 3,910 operational taxonomic units (OTUs) based on the basis of a 97% sequence identity threshold. At the phylum level, sequencing of all samples revealed the assignment of reads to 26 distinct microbial phyla. The gut microbiome was dominated by the phyla Firmicutes and Bacteroidota across both seasons and locations. The average relative abundances (± SD) of Firmicutes were as follows: Lampang_wet 44.71 ± 2.88, Lampang_dry 50.43 ± 6.31, Kanchanaburi_wet 53.76 ± 5.76, and Kanchanaburi_dry 54.07 ± 5.81. For Bacteroidota, the corresponding values were as follows: Lampang_wet 33.75 ± 2.84, Lampang_dry 29.45 ± 5.39, Kanchanaburi_wet 25.26 ± 3.44, and Kanchanaburi_dry 23.33 ± 8.73. These were followed in abundance by Spirochaetota, Verrucomicrobiota, and Proteobacteria (Fig. 1a). The order Bacteroidales was the most abundant order across all the groups—Lampang_wet, Lampang_dry, Kanchanaburi_wet, and Kanchanaburi_dry—with relative abundances of 33.31 ± 2.74, 26.64 ± 8.77, 24.89 ± 3.22, and 21.91 ± 8.52, respectively. This was followed by Lachnospirales (15.51 ± 3.19, 11.99 ± 3.65, 17.40 ± 3.28, and 8.02 ± 2.55), Oscillospirales (10.78 ± 1.58, 10.50 ± 4.39, 14.15 ± 4.30, and 7.56 ± 2.44), Bacillales (0.16 ± 2.44, 11.43 ± 15.34, 0.16 ± 0.05, and 21.29 ± 12.12), and Spirochaetales (8.09 ± 2.40, 5.60 ± 1.92, 7.37 ± 4.35, and 2.61 ± 1.37), respectively (Fig. 1b). The 5 most abundant families were Lachnospiraceae (15.29 ± 3.18, 11.81 ± 3.63, 11.26 ± 3.63, and 7.93 ± 2.54, respectively), Rikenellaceae (10.58 ± 1.83, 8.95 ± 3.30, 7.94 ± 2.40, and 8.52 ± 2.11), Spirochaetaceae (8.09 ± 2.40, 5.60 ± 1.92, 7.37 ± 4.35, and 2.61 ± 1.37), Oscillospiraceae (6.25 ± 0.95, 6.34 ± 2.61, 7.52 ± 2.05, and 4.17 ± 0.88), and Planococcaceae (0.08 ± 0.08, 10.34 ± 14.00, 0.12 ± 0.04, and 17.39 ± 11.60), respectively (Fig. 1c). Rikenellaceae_RC9_gut_group was the most abundant genus (9.41 ± 1.85, 8.26 ± 3.04, 6.69 ± 1.89, and 8.03 ± 2.15), followed by f__Lachnospiraceae_Unclassified (7.61 ± 1.64, 5.85 ± 1.71, 9.43 ± 2.00, and 4.30 ± 1.44), Treponema (7.75 ± 2.46, 5.32 ± 1.93, 7.00 ± 4.55, and 2.34 ± 1.38), p-251-o5 (6.01 ± 1.49, 3.15 ± 1.58, 3.71 ± 2.38, and 1.37 ± 1.28), and WCHB1-41 (4.00 ± 1.47, 4.33 ± 2.17,3.44 ± 2.23, and 3.44 ± 1.94), respectively (Fig. 1d).

The diversity of the gut microbiome

As shown in the rarefaction curves of the OTUs (Supplementary Fig. 1), the observation that the OTU rarefaction curves neared a plateau suggested that the majority of microbial diversity within each sample was likely accounted for. The alpha-diversity indices, including Chao1 (species richness) and Shannon (species diversity), indicated that Lampang_wet group exhibited the highest diversity (Chao1: 2397.60 ± 92.60; Shannon: 8.96 ± 0.18), whereas the Kanchanaburi_dry group displayed the lowest diversity (Chao1: 1992.92 ± 196.88; Shannon: 7.40 ± 0.79). Furthermore, significant differences were found across the 4 groups for both indices using the Wilcoxon rank-sum test (p = 0.0004 and p = 0.00036, respectively) (Fig. 2). Specifically, the diversity values were significantly higher in the wet season than in the dry season and in Lampang than in Kanchanaburi (Fig. 3). Furthermore, beta-diversity analysis (community dissimilarity among different groups), visualized using Principal Coordinates Analysis (PCoA) based on the Bray-Curtis distance, revealed a significant distinction in community structure between the groups (Fig. 4a). Additionally, ANOSIM analysis showed a significant difference between groups (R = 0.409, p = 0.001) (Fig. 4b).

Differences in the gut microbiome

Many diverse microbial taxa were highly abundant across different seasons and between Lampang and Kanchanaburi region. The top 10 genera with the highest absolute abundance that were significantly different between samples from Kanchanaburi and Lampang, as well as between seasons, were identified using the Wilcoxon Rank-Sum test (Fig. 5). Specifically, Saccharofermentans was most abundant in the Lampang group, while p_251_o5 and Fibrobacter were more abundant in the wet season than in the dry season, which is consistent with the differences in proportional abundance observed from the STAMP test (Supplementary Figs. 2 and 3). From the LEfSe analysis, significant variations in abundance were observed across locations (Fig. 6a), seasons (Fig. 6b), and interactions between location and season (Fig. 6c). The phyla Bacteroidota, Spirochaetota, and Fibrobacterota, and the genera Treponema, Fibrobacter, p_251_o5, and Prevotellaceae_UGC_003 were elevated in the Lampang samples. In Kanchanaburi, the phylum Firmicutes and the genera Clostridium sensu stricto_3, Lysinibacillus, and Solibacillus were significantly more abundant. In wet season samples from Lampang, the orders Bacteroidales, Spirochaetales, Fibrobacterales, and Acidaminococcales were highly abundant. In contrast, Flavobacteriales, Weeksellaceae, and Paludibacteraceae were uniquely enriched in dry season samples from Lampang. In Kanchanaburi, the wet season was characterized by the high abundance of several gut microbiome members, including the orders Lachnospirales, Oscillospirales, Christensenellales, Erysipelotrichales, and Synergistales; families such as Oscillospiraceae, Erysipelotrichaceae, and Synergistaceae; and genera including Sarcina and Eubacterium coprostanoligenes. Notably, the order Enterobacterales, including Escherichia_Shigella and Citrobacter, ​​Bacillales, and Pseudomonadales were particularly observed in the Kanchanaburi samples during the dry season.

Functional differences in the gut microbiome of elephants

The main classifications of microbial functions were metabolism, genetic information processing, and environmental information processing (Fig. 7a). A greater number of distinct functional pathways were identified between seasonal groups than between locations (data not shown). Metabolic pathways associated with food digestion and processing—such as glycan biosynthesis and metabolism, arginine and proline metabolism, and glycine, serine, and threonine metabolism—were enriched in samples from Lampang. In contrast, transporter pathways, including transporters and membrane transport, were enriched in samples from Kanchanaburi. The most notable difference was the enrichment of carbohydrate metabolism pathways, including carbohydrate metabolism, starch and sucrose metabolism, fructose and mannose metabolism, galactose metabolism, pentose and glucuronate interconversions, pentose phosphate pathway, methane metabolism, and carbon fixation in photosynthetic organisms, which were observed exclusively in the wet season. Conversely, energy metabolism pathways (citrate cycle [TCA cycle], pyruvate metabolism, synthesis and degradation of ketone bodies, butanoate metabolism, and propanoate metabolism) and cofactor and vitamin metabolism pathways (ascorbate and aldarate metabolism, riboflavin metabolism, retinol metabolism, and folate biosynthesis) were more prominent in the dry season than in the wet season (Fig. 7b).

This study provides novel insights into the spatiotemporal dynamics of the gut microbiome in captive Asian elephants in Thailand, demonstrating that both seasonal and location factors significantly influence microbial composition, diversity, and predicted functional profiles. These findings support the hypotheses that environmental factors and dietary shifts, even under semi-controlled captive conditions, shape the gut microbial communities of large herbivores such as elephants. Consistent with previous reports in hindgut fermenters and other megaherbivores, including Asian elephants, the gut microbiome was dominated by the phyla Firmicutes and Bacteroidota^4,7,12,13. Firmicutes are well known for their roles in fiber degradation and short-chain fatty acid production¹⁴, while Bacteroidota breaks down complex carbohydrates, including recalcitrant dietary glycans, through specialized polysaccharide utilization loci (PULs) that enhance their ability to process diverse plant materials¹⁵.

Beyond their overall dominance, the relative abundances of Firmicutes and Bacteroidota displayed distinct seasonal and locational patterns that likely reflect differences in dietary fiber availability and energy-harvesting strategies. Our results revealed that Firmicutes proportions were higher in Kanchanaburi and during the dry season, whereas Bacteroidota were more abundant in Lampang and during the wet season. This pattern aligns with the functional roles of these phyla: Firmicutes are associated with diets high in structural plant fiber and are responsible for degrading cellulose into volatile fatty acids that support energy harvest¹⁴, while Bacteroidota specializes in breaking down non-fibrous polysaccharides that enable efficient processing of diverse plant materials¹⁵.

At the genus level, key Firmicutes taxa enriched in Kanchanaburi and during the dry season included Clostridium sensu stricto, Lysinibacillus, and Solibacillus, which have been associated with fiber degradation and adaptation to dietary restrictions^16,17,18. Members of Lachnospiaceae, a major Firmicutes family, are known for producing short-chain fatty acids (SCFAs) including butyrate, propionate, and acetate through fermentation of dietary fiber¹⁹, thereby enhancing energy extraction under limited forage conditions. By contrast, Bacteroidota-affiliated taxa and fibrolytic genera such as Fibrobacter, Saccharofermentans, Provotella, were enriched in Lampang and during the wet season, when elephants had access to more diverse forage including natural forest vegetation. These genera are specialized in carbohydrate and hemicellulose breakdown²⁰ and their enrichment coincided with upregulated carbohydrate metabolism pathways (including starch, sucrose, fructose, and mannose metabolism) observed exclusively in the wet season. Taken together, these seasonal and locational shifts in the Firmicutes–Bacteroidota ratio and their associated key genera suggest that the elephant gut microbiome dynamically adjusts its fiber-degradation and energy-harvesting capacity in response to changes in dietary diversity and resource availability across sites and seasons.

The findings of this study revealed that the diversity of the gut microbiome in Asian elephants undergoes significant seasonal changes, with microbial diversity being higher during the wet season than during the dry season. This pattern is consistent with observations in other large herbivores such as forest musk deer, plateau pikas, and geladas^21,22, where microbial richness increases during periods of enhanced dietary complexity or fiber availability. In this study, different seasonal genus markers were identified, with Fibrobacter, a fibrolytic and cellulolytic bacterium associated with plant digestion, being characteristic of the wet season, and Lysinibacillus and Solibacillus being more prevalent during the dry season. In addition, Saccharofermentans, fibrolytic and cellulolytic bacteria, as well as Prevotella, a fermentative genus, were more abundant during the wet season than during the dry season, particularly in Lampang. This finding is supported by a reported negative correlation between Fibrobacter with Lysinibacillus and Solibacillus in the elephant gut across different populations in China, which is associated with dietary characteristics²³. Lysinibacillus has been shown to contribute significantly to changes in gut microbial diversity in response to environmental factors such as diet and drinking water²⁴, while Solibacillus has been found in high abundance in conditions possibly linked to food scarcity in elephants^4,25. Thus, these bacterial taxa may be useful indicators for assessing food sufficiency and seasonal variation in the gut microbiota of elephants. The observed seasonal differences in microbial diversity and composition are likely linked to changes in forage availability between the wet and dry seasons. In northern Thailand, natural vegetation becomes more abundant and diverse during the wet season, potentially increasing the range of plant material available to elephants with forest access, such as those at Lampang, and supporting the enrichment of fibrolytic taxa like Fibrobacter and Saccharofermentarus. Similarly, elephants in Kanchanaburi had limited forest access during the wet season but relied almost entirely on cultivated forage in the dry season. However, we did not systematically quantify individual dietary intake or plant species selection across seasons, so we cannot definitively attribute microbiome changes to specific dietary components. This limitation should be considered when interpreting our findings, and future work incorporating direct dietary monitoring (e.g., fecal botanical analysis or DNA metabarcoding) would strengthen understanding of diet–microbiome relationships in captive elephants.

The findings also revealed that location significantly influenced the composition of the gut microbiome independent of seasonal variation. Compared with those in Kanchanaburi (western Thailand), the elephants in Lampang (northern Thailand) exhibited greater microbial diversity and a distinctly different community structure compared to those from Kanchanaburi (western Thailand), possibly resulting from the wider variety of food available in Lampang. Members of the phyla Bacteroidota, Spirochetota, and Fibrobacteriota were highly abundant in Lampang, whereas the phylum Firmicutes, which is primarily represented by Clostridiales, was more abundant in Kanchanaburi. Spirochetota and Fibrobacteriota have been reported to be enriched in captive Asian elephants, which are typically fed Napier grass, whereas Firmicutes tends to be more abundant in wild Asian elephants that consume lower-fiber crops and grains²⁶. In contrast, our study involved elephants in Lampang that were fed a variety of grass types, while those in Kanchanaburi were primarily fed Napier grass and pineapple trunks. These differences may result from regional variations in management practices, vegetation, soil microbiota^27,28,29. Additionally, anthropogenic factors such as human interaction, tourist feeding, translocation, deworming, and captive conditions^24,30 may influence the gut microbiome of conservation- and tourism-managed elephants in Lampang and Kanchanaburi.

Enterobacterales, including Escherichia–Shigella and Citrobacter, were particularly abundant in samples from Kanchanaburi during the dry season. Commensal strains of Enterobacterales have been identified in the gut microbiomes of elephants^4,26,31. However, pathogenic strains, which are of public health concern due to their association with diarrhea and antimicrobial resistance, may be transmitted through contact with elephants, humans, the diet, or the environment⁴. Escherichia–Shigella has shown dramatic shifts in abundance across seasonal transitions in red pandas and has been associated with gastrointestinal distress in giant pandas³². The high abundance of Enterobacterales in elephants and its potential relationship with gut health warrant further investigation. Several ecological factors may have contributed to the observed Enterobacterales bloom in Kanchanaburi during the dry season. First, elephants in Kanchanaburi were primarily fed a monotonous diet of Napier grass and pineapple trunks with restricted or entirely prohibited forest access during the dry season, resulting in limited dietary diversity compared to the more varied forage available in Lampang. Studies in other herbivores have shown that low-fiber or monotonous diets can trigger substantial declines in gut bacterial diversity and promote expansion of Proteobacteria, including Enteobacteriaceae³³. Second, the dry season in western Thailand is characterized by reduced water availability and increased environmental temperatures, which may impose physiological stress on captive elephants and alter gut conditions that favor facultative anaerobes such as Enterobacterales³⁴. Third, management practices including higher tourism intensity, restricted movement, and limited access to natural water sources during the dry season may contribute to environmental stress and microbiome perturbation. These converging stressors—dietary restriction, water scarcity, and management pressure—are consistent with patterns observed in other captive and wildlife populations, where environmental stress correlates with reduced microbiota diversity and increased abundance of stress-tolerant taxa such as Enterobacteriaceae³⁵.

In this study, functional prediction analysis revealed that the unique relative abundances of microbial taxa were more strongly associated with seasonal variation than with location. The most notable difference was the enrichment of carbohydrate metabolism pathways such as general carbohydrate metabolism, starch and sucrose metabolism, and fructose and mannose metabolism which were observed exclusively during the wet season. The captive elephant diet, which includes a high proportion of carbohydrates, particularly during the wet season, requires increased cellulolytic bacterial activity for digestion. Several bacterial taxa enriched in the wet season, including Fibrobacteriota, Saccharofermentans, Spirochaetaceae, and Clostridia, play key roles in cellulose degradation and nutrient absorption in herbivores^24,26. It is important to note that our analysis is based on PICRUSt predictions derived from 16 S rRNA gene data and therefore reflects inferred rather than directly measured microbial functions. PICRUSt relies on reference genomes and phylogenetic relatedness to estimate gene content, which can introduce bias when the studied community contains taxa that are under-represented or absent in reference databases. Thus, the predictions should be interpreted with appropriate caution and ideally validated in future work using shotgun metagenomics, metatranscriptomics, and metabolomics to directly quantify functional genes and metabolites in captive Asian elephants.

Study sites and sample collection

The study included healthy elephants from two locations in Thailand: 20 individuals (5 males, 15 females; age range from 7 to 65 years old) from the Thai Elephant Conservation Center, Lampang (18.3601° N, 99.2472° E), in the northern region, and 10 individuals (5 females; age range from 16 to 60 years old) from a private elephant camp in Kanchanaburi (14.2172° N, 99.2242° E), in the western region (Fig. 8a). Although sample sizes per site and season were not identical due to the limited number of the elephants, all samples were collected and processed using an identical protocol, and sequencing depth was normalized by rarefaction to a common read count prior to diversity and community-level analyses, which reduces technical bias associated with unequal sequencing effort. At the Thai Elephant Conservation Center, Lampang, two distinct elephant management practices were observed. (1) Non-Forest-Based Management: Elephants in this group were maintained in permanent captivity and provided a consistent, year-round diet. This diet, procured from local farmers, primarily consisted of fresh grasses, including Pennisetum x P. americanum hybrid (Bana grass), Digitaria eriantha (pangola grass), and Brachiaria mutica (para grass), supplemented with dry Pangola hay. Additional food items included bananas, sugarcane, tamarind, and seasonal fruits. (2) Forest-Access Management: Elephants in this group received the same cultivated diet as those in the non-forest-based group, but were also permitted to forage in natural forest areas daily, typically from 08:00 to 15:00. Foraged items encompassed bamboo species, naturally occurring grasses, and foliage from trees such as Albizia saman (rain tree) and Ficus spp. (fig tree). The locations within the forested areas were varied. During the night, these elephants were provided with food. At the elephant camp in Kanchanaburi, elephants were primarily fed cultivated forage, including Pennisetum purpureum (Napier grass) and pineapple trunks. During the wet season, elephants were granted limited forest access (approximately 1–2 h per day), where they consumed native vegetation such as Thyrsostachys siamensis (Long sheath bamboo) and various vines. However, during the dry seasons, forest access was either restricted or entirely prohibited, and elephants relied predominantly on cultivated forage. Nevertheless, we did not systematically record individual dietary composition. All diet represents the routine feeding regimen for the animals throughout the year.

The fecal samples (5–10 g) were freshly collected from a total of 30 healthy adult captive elephants in two distinct seasons: the wet season (September/August 2024) and the dry season (December 2024/January 2025) under the permission from both Thai Elephant Conservation Center and a private elephant camp (Fig. 8b). In this study, fecal samples were analyzed at the individual level; samples were not pooled across elephants within a group, site, or season. Each elephant, therefore, contributed one microbiome profile per sampling season. Data were retrieved from ClimateCharts.net to represent local climate conditions, including wet and dry seasons. The chart was generated using the Walter-Lieth climate diagram, which visualizes temperature and precipitation relationships to highlight periods of dry and wet months³⁶. Each sample was collected using a sterile glove and immediately placed into a 50 mL sterile tube containing DNA/RNA Shield (Zymo Research, Irvine, CA, USA). Samples were transported at environmental temperature and subsequently stored at −20 °C at the molecular laboratory, Faculty of Veterinary Science, Mahidol University, until further analysis.

All experiments were performed in accordance with relevant guidelines and regulations under the Institutional Animal Care and Use Committee (IACUC), Faculty of Veterinary Science, Mahidol University under the COA-IACUC number MUVS-2022-12-68.

DNA extraction

DNA from the fecal samples was extracted using the ZymoBIOMICS^® DNA Miniprep kit (ZymoResearch, Irvine, CA, USA) following the manufacturer’s recommendation. After extraction, the DNA was qualified and quantified using a NanoDrop™ One spectrophotometer (Thermo Fisher Scientific, USA). Finally, the quality of the extracted DNA was additionally assessed using gel electrophoresis.

Library preparation and sequencing

The sequencing library was constructed using a MetaVX Library Preparation Kit (Beijing Baimaike Biotechnology Co., Ltd., Nanjing, China). Then, generate amplicons that cover V3-V4 hypervariable regions of the 16 S rRNA gene of bacteria. The V3–V4 region contains enough sequence variation to characterize microbial communities at higher taxonomic ranks such as phylum and genus, and it has become a commonly selected target for studies of microbiota diversity³⁷. The forward primer contains the sequence 5′-barcode ACTCCTACGGGAGGCAGCAG-3′ and the reverse primers contains the sequence 5′-barcode- GGACTACHVGGGTWTCTAAT-3′, were generated³⁷. The sequencing was conducted on an Illumina Novaseq platform (Illumina, San Diego, USA).

Bioinformatic analysis of the microbiome

The raw sequence data were processed for quality control by removing the sequence of N, retains the sequence length was > 200 bp. After quality filtering and removing chimeric sequences, Operational Taxonomic Units (OTUs) were clustered using VSEARCH v1.9.6 at a 97% sequence similarity threshold. The representative sequences for each OTU were then taxonomically classified using the RDP Classifier v 2.2 with a Bayesian algorithm, referencing the SILVA 138 16 S rRNA gene database³⁸. To characterize community composition, OTU-level taxonomic assignments were summarized at various taxonomic ranks (e.g., phylum, class, genus) for each sample. KronaTools v 2.7³⁹ software was used to visualize the distribution of different species on different levels. For diversity analyses, the OTU table was normalized by random subsampling (rarefaction) to equalize the sequencing depth across samples. Alpha diversity metrics, including the Shannon and Chao1 indices, were calculated to assess species diversity and richness, respectively, using Qiime v 1.9.1⁴⁰. In addition, rarefaction curves and rank-abundance plots were generated to evaluate species richness and community evenness. Beta diversity was assessed via Principal Coordinates Analysis (PCoA) based on based on Bray–Curtis dissimilarity matrices. For multi-group comparisons, Linear Discriminant Analysis Effect Size (LEfSe) v 1.0⁴¹ was used to identify taxa that significantly differed across groups. PICRUSt v 1.0⁴² was used to predict the functional potential of microbial communities based on the basis of 16 S rRNA gene data. The functional annotations were inferred using the Kyoto Encyclopedia of Genes and Genomes (KEGG) Orthology (KO) databases⁴³⁴⁴., which enables the prediction of gene content and metabolic pathways present in each sample.

Statistical analysis

For alpha diversity analyses, non-parametric tests were applied. For comparisons between more than two groups (four groups: combinations of location and season), the Kruskal–Wallis test was used to evaluate overall differences in alpha diversity indices. When significant, post hoc pairwise comparisons between groups were performed using the Wilcoxon rank-sum test. For two-group comparisons, the Wilcoxon rank-sum test was applied directly. To control for multiple testing in pairwise comparisons, p-values were adjusted using the Holm correction. A corrected p < 0.05 was considered statistically significant.