Tropical mountains cover less than 10% of the terrestrial land but support approximately half of the global biodiversity hotspots (Myers et al., 2000; Mittermeier et al., 2011; Quintero and Jetz, 2018). Various theories have been proposed to explain the extraordinarily high biodiversity in the tropics, yet a large number of species in tropical mountains remain one of the ecology’s unsolved puzzles (Mittelbach et al., 2007; Fine, 2015). Janzen (1967) proposed that mountain climates played a central role in determining the biodiversity in tropical ecosystems, which is widely known as the mountain passes hypothesis (Ghalambor et al., 2006). Focusing on seasonal variations, Janzen (1967) suggested that low climatic variability across climatic gradients creates climatic barriers for species, leading to population divergence and eventually high richness in tropical mountains. Several studies have confirmed various aspects of Janzen’s hypothesis with respect to plants and animals (Eo et al., 2008; McCain, 2009; Polato et al., 2018). However, it has remained unclear whether soil bacteria and fungi, key regulators of terrestrial biogeochemical processes, also display strong climatic variability-related patterns along different elevations in tropical montane systems.

Over the past decade, several studies have been carried out to characterize the shift in soil microbial communities along montane gradients, although tropical regions have been less examined (Hendershot et al., 2017; Looby and Martin, 2020). In these previous studies, specific groups of microbes often have unique relationships with elevation (Kivlin et al., 2017; Looby and Martin, 2020), and furthermore, community differences were best explained by elevation-related changes in soil properties, such as pH and carbon or nitrogen content, and usually not by climatic parameters per se (Bryant et al., 2008; Dai et al., 2021). Only a few studies along elevational transects revealed strong correlations between microbial distribution and climatic parameters (Nottingham et al., 2018a; Frindte et al., 2019; Geml et al., 2021; Ma et al., 2022). One potential explanation for the lack of a consistent climatic relationship is that most prior studies rely on low-resolution macroclimate data. For instance, mean annual temperature derived from synoptic weather stations in combination with grid-based interpolations are often used to explain diversity patterns in microbial communities, although they are known to be different from the soil climates that microbes experience (Li et al., 2013). Importantly, climatic variability such as temperature extremes within and outside the growing season can differentially affect different taxonomic groups or stages of microbial community development (Bardgett and Caruso, 2020; Ma et al., 2022). Moreover, tropical mountain climates have variability in seasonal rainfall and soil water availability (Sarmiento, 1986). These climatic features could shape microbial communities by directly mediating metabolic rates and indirectly affecting resource availability via controlling aboveground productivity and inputs of carbon and nitrogen to soils (Belnap et al., 2005; Körner, 2007; Keitt et al., 2016). To test Janzen’s mountain passes hypothesis for soil microbes in tropical montane systems, fine-scale climate data that reflect the climatic differences between soil cores and the variability of the climatic parameters are needed.

Tropical montane ecosystems are further characterized by frequent cloud immersion at the mid to upper elevations, which affects the climatic conditions (Bruijnzeel et al., 2011), and therefore likely regulates microbial diversity and community composition. Frequent cloud cover contributes to patterns in forest biodiversity (Hietz, 2010; Karger et al., 2021), by influencing moisture and temperature and modulating soil properties such as carbon and nitrogen content (Zimmermann et al., 2009; Fahey et al., 2016; Hernández-Vargas et al., 2019). Thus, we expect microbial communities to respond to climatic patterns. Studies on the impacts of such unique climatic phenomena on soil microbes are currently scarce, while some studies focused mainly on plant diversity patterns (Oosterhoorn and Kappelle, 2000; Fahey et al., 2016) or specific microbial groups such as ectomycorrhizal communities associated with cloud forest plants (Morris et al., 2009; Olmo-Ruiz et al., 2017).

In this study, we tested Janzen’s mountain passes hypothesis along a ~1,200 m tropical elevational gradient on Hainan Island of South China, exploring the patterns and importance of climatic variability in spatial variations of soil bacterial and fungal communities. Given the steep environmental gradient associated with increasing elevation, we expected marked shifts and directional changes in the diversity and composition of soil bacterial and fungal communities with changes in elevation (Callaway et al., 2002). We used climatic data in the soil layer from each site to calculate local-scale climatic variability, including the range of soil temperature and moisture in the growing season and the seasonality of soil temperature and moisture, among others. We expected that the local-scale soil climatic variables would play more critical roles than soil chemistry and vegetation in shaping bacterial and fungal diversity patterns and composition along the tropical elevational gradient, thereby supporting the Janzen’s hypothesis (Frindte et al., 2019; Ma et al., 2022). Likewise, we expected that soil microbial communities would respond to the unique climate in cloud forests and hold a unique pattern in diversity and composition compared with other forests in the mountain (Zimmermann et al., 2009; Velez et al., 2021).

Our first expectation that the microbial diversity and composition display directional changes along the tropical montane gradient was verified. We revealed that both bacterial and fungal diversity, measured as richness and Shannon index, declined as elevation increased, and the dissimilarity of bacterial and fungal communities increased with increased elevation differences (Figures 1, 2). While geographical patterns of soil microbial communities have been a hot topic in recent years, few studies have examined their distribution along elevational gradients in the tropics (Wang et al., 2022). Our results demonstrated a significant role for the abiotic environment in shaping the diversity and composition of soil bacteria and fungi along the elevational gradient. Fundamentally, climatic factors, including soil temperature and moisture and certain edaphic factors, particularly pH, were strongly associated in our data with variations in bacterial and fungal diversity and composition along the montane gradient.

Our results generally supported the Janzen’s hypothesis. We confirmed that climatic variability had a strong influence on the spatial variation of soil bacterial and fungal composition and fungal diversity along a tropical montane elevational gradient. Although effects of climatic variability on microbial diversity and composition have been observed in ecosystems such as forests and grasslands (Rasche et al., 2011; Shi et al., 2020), previous studies have often found soil properties such as pH and carbon or nitrogen content as the dominant factors that explained microbial distribution along mountainsides (Singh et al., 2012; Shen et al., 2013; Zhang et al., 2015). Using local-scale soil temperature and moisture data instead of macroclimate data might be the key to resolving this contradiction. Meanwhile, climatic differences between soil cores reflect potential climate extremes that determine microbial survival and climatic variability that meditate compositional shifts in microbial communities directly and indirectly (Bell et al., 2009; Keitt et al., 2016). Specifically, the range of soil moisture in the growing season and moisture seasonality were the main drivers of fungal diversity along our elevational gradient. There is evidence indicating that fungi, particularly fungal spores, are more sensitive than bacteria to changes in soil moisture (Griffin, 1963; Kaisermann et al., 2015), and in some cases fungal abundance based on qPCR of 18S rRNA gene was more reduced by drought (Cregger et al., 2012). Because the daily or circadian climate cycle is markedly higher than the differences in climate from month to month in tropical regions (Janzen, 1967; Sarmiento, 1986), the negative correlation between fungal diversity and the range of moisture during the growing season suggests that dry soil conditions negatively influence fungal diversity (He et al., 2017; Preece et al., 2019), whereas the positive relationship with seasonal variation in soil moisture suggests niche partitioning with respect to soil moisture could be facilitating fungal diversity (Kaisermann et al., 2015; Větrovský et al., 2019). In addition, soil temperature directly affects microbial metabolism and the seasonal changes are closely linked to short- and medium-term variations in the quantity and quality of resources entering the soil (Krave et al., 2002; Cookson et al., 2006; Bell et al., 2009), which further correlated with compositional variation of both bacterial and fungal communities.

Besides climatic attributes, soil properties as pH, NH₄⁺, TP, water content, and plant evenness and tree biomass storage were also crucial for shaping microbial diversity and community composition. Consistently, soil pH has been documented as the key driver of bacterial communities along elevational gradients globally (Wang et al., 2022). It has also been documented that increased nitrogen limitation could lead to more fungi-dominated microbial communities at higher elevations (Fierer et al., 2009; Nottingham et al., 2018b). Generally, microbial composition and abundance depend on soil nutrient availability (Lozupone and Knight, 2007; Jesus et al., 2009). Meanwhile, plant attributes of diversity and biomass (simplified as diameter at breast height of trees) may regulate bacterial and fungal communities by determining the quantity and quality of the litter and root exudate supply (such as C or N source) and by modifying the soil physical environment (Wallenstein et al., 2007). Particularly, a more even plant community should result in a more even distribution of roots. There is evidence that plants can alter root growth strategies in response to variations in their neighbors and soil nutrients (Wambsganss et al., 2021). An even distribution of roots, thus, should maximize the volume of the soil explored by plant roots and maximize the number of microhabitats available to the soil microbial community.

Biodiversity in cloud forests is still largely under-documented and poorly understood (Karger et al., 2021), particularly soil microbial communities. Our data revealed a more diverse bacterial community and divergent composition of both bacterial and fungal communities in the cloud forests (Figure 3A,B, 4A–D; Supplementary Figure 2). The findings are consistent with our expectations and suggest that the microbial community, especially bacteria in the cloud forests, was substantially different from those of other forests along the montane gradient. Notably, humid cloud forest soils with less fluctuation in temperature (decreased seasonality of soil temperature and increased mean annual soil moisture in Supplementary Table 2) may foster unique microbial communities. On the one hand, the relatively stable environments, compared to other elevational ranges, may enable increased survival and even speciation in microbial taxa, especially in bacterial communities (Pellissier, 2015), which deserve further investigation. On the other hand, microbial life-history strategies may also explain part of the phenomenon. For instance, phyla of Acidobacteria prefer soil with high moisture, whereas Chloroflexi prefer warmer soil conditions (Jones et al., 2009; Oliverio et al., 2017; Pérez Castro et al., 2019). Hence, Acidobacteria may be particularly favored in the cloud forest soils, but Chloroflexi was reduced (Supplementary Table 4).

As tropical montane ecosystems may face considerable threats from future climate change (Pounds et al., 1999; Foster, 2001), disentangling the role of climate in the distribution of biotic communities along tropical montane gradients is highly valuable for predicting the consequences of climate change. Here, we show that soil bacterial and fungal communities on tropical montane displayed monotonic changes with increasing elevation and that bacterial and fungal diversity and community composition were largely predicted by climatic variability. Since global warming is expected to increase seasonal variability and alter daily temperature extremes (Karl et al., 1991; Wang and Dillon, 2014), climate change is likely to shift microbial communities in tropical montane systems and impact terrestrial biogeochemical cycling, particularly in cloud forests.

Finally, we acknowledge that the results reported here come from only one gradient study. Tropical montane systems are highly variable in climatic conditions due to differences in topography, cloud climatology, and geographic location (Janzen, 1967; Sarmiento, 1986). This climatic variability could lead to divergent patterns in soil microbial diversity and composition but is also a crucial ecological characteristic for life and survival of microorganisms. As such, similar studies in other tropical elevational gradients are needed for a more unified understanding of mountain microbes. Our work shows how the climatic variability perspective can provide new insights into microbial ecological studies.

The data presented in the study are deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) repository, accession number PRJNA903527.

LL, YD, and WL contributed to the conception and design of the study. YF, JW, XQ, and LL participated in the sample collection and processing. YF, JZ, YD, and LL contributed to the data analysis, methodology, visualization, and validation. YF wrote the original draft. LL organized the field expedition and acquired the funding. All authors contributed to the review, editing, and approval of the final submitted version.

This research was financially supported by Hainan Natural Science Fund (322RC581 to LL), Hainan University (KYQD(ZR)21116 to LL), and Fundamental Research Funds for the Central Non-profit Research Institution of Chinese Academy of Forestry (CAFYBB2022SY024 to YD).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.