Environmental drivers
The effect of elevation on fungal community composition and richness is mediated through abiotic and biotic factors driven either directly or indirectly by differences in temperature, which strongly influences relative humidity, soil moisture, and soil chemical processes. Consequently, the forest types found in different elevation zones have distinct mesoclimatic and edaphic conditions. Moreover, higher elevation habitats generally cover smaller areas than lower elevation ones and habitat area often is correlated positively with species richness in many taxonomic groups (Gotelli 1998). However, for fungi, the decreasing habitat area with increasing elevation does not seem to offer a satisfying explanation for the observed richness patterns: there were no statistical difference in total fungal richness among lowland and upper montane forests in any of the three regions, even though lowland forests cover much larger areas, particularly in Borneo and Panama (Figs. S1-3). Although some functional groups decreased in richness with increasing elevation, other groups showed the opposite trend. It is likely that climatic and edaphic factors as well as composition of biological communities are far more influential drivers of fungal richness in mountains than habitat area alone.
Our results indicate that MAT is the most influential driver of fungal community composition (Figs. 1-2, Table 1), while the role of MAP seems to be more region-dependent. For example, in terms of community composition, MAP was the strongest contributor of explained variation in the majority of the functional groups in Argentina, while it remained marginal or insignificant in Borneo and Panama. The mainly subtropical Yungas has pronounced seasonality, particularly at lower elevations (Brown et al . 2001). Local differences in moisture, and the resulting temporary drought stress, may represent a stronger environmental filter for community assembly than the limited variability and greater amounts of available moisture in the wet tropical forests of Borneo and Panama, which have far less pronounced seasonality.
Soil pH also plays an important role in shaping belowground fungal communities (Porter et al. 1987; Coughlan et al. 2000; Lauber et al. 2008; Rousk et al. 2010; Geml et al . 2014; Tedersoo et al . 2014; Glassman et al. 2017, Geml 2019). Because many fungal species have a relatively wide pH optimum (e.g., Wheeler et al. 1991; Nevarez et al. 2009), it is likely that the observed correlation of pH with community composition is mainly indirect, e.g., by altering nutrient availability and competitive interactions between soil fungi and bacteria (Rousk et al. 2010) and other soil biota. In all regions, we observed significant decrease in pH and increase in OM with elevation, and it is difficult to disentangle their effects from that of MAT. However, the observed increase in pH and decrease in OM in the sparsely vegetated granitic summit zone of Mt. Kinabalu suggest that within the forest habitat, lower temperatures are associated with greater OM accumulation and a simultaneous decrease in pH in the montane forests, whereas OM accumulation drops sharply in sparsely vegetated habitats near the summit and does not result in low soil pH.