Pantropical core mycobiome of lowland and montane forests
The rarefied, combined dataset included 7612 OTUs. Of these, only 211 OTUs were shared among all three regions. The majority of OTUs were restricted to one geographic region, but 376 were shared between the two Neotropical regions, 188 were shared between the Yungas and Borneo, and 192 were shared between Borneo and Panama (Fig. 4). These findings are consistent with Hypothesis 3, i.e., little overlap among species pools of the three biogeographic regions, with more shared species between the two Neotropical regions.
Both abiotic environmental variables and geographic location were strongly correlated with fungal community structure in separate Mantel tests (r = 0.658 and r = 0.497, respectively; all p< 0.001), and when they were combined in partial Mantel tests, resulting in significant correlation of environmental factors with fungal community structure when location was accounted for (r = 0.553, p < 0.001), and vice versa (r = 0.279, p < 0.001). PerMANOVA indicated that geographic region and forest type explained 7.3% and 5.8% of the variation in community composition, respectively, while among abiotic variables, MAP explained 7.4%, pH 6.9%, and MAT 5.7% of the variation, all with significant contributions in the combined model, consistent with Hypothesis 1.
Of 360 fungal OTUs with significant (p < 0.05) indicator value for a forest type, 136 occurred both in the Neo- and Paleotropics and were included in the pantropical comparison. Of these, 40, 24, and 72 were indicators for the lowland, lower montane, and upper montane forests, respectively (Table S2).
Discussion
Fungal biodiversity in tropical forests remains little known, and opportunities to compare data from similar guilds across diverse tropical forest types at local and global scales are rare. The deep sequence data presented here show that composition of the total fungal community, as well as that of all functional groups, is strongly structured according to elevational forest types in both the Neo- and the Paleotropics. Contrary to vascular plants, where the lowland and the lower montane forests typically harbor more species than upper montane forests (Aiba and Kitayama 1999; Brown et al. 2001; McCain and Grytnes 2010), we did not find substantial differences in soil fungal richness among the three elevational zones in Argentina and Borneo. The lack of a strong elevational pattern in fungal richness is similar to the lack of latitudinal differences in fungal richness on a global scale (Větrovský et al. 2019). Panama was an exception, where the mid-elevation peak in fungal richness is concordant with vascular plant richness in Central American mountains (Cardelús et al. 2006, Prada et al. 2017). However, in all functional groups and in all regions, compositional structure appears to be driven by elevation and the resulting environmental filtering according to contrasting climatic and edaphic conditions, and to a smaller extent, differences in vegetation. Furthermore, composition of fungal communities in the lower montane forests may be regarded as intermediate between communities of the lowland and upper montane forests, although, as the other two forest types, the lower montane forest also possesses several characteristic taxa.