Spatial differences
Different Amazonian habitats varied considerably in their biotic composition10, 65. Habitat was the most significant factor explaining community turnover in the 18S and COI datasets. Although habitat was also significant for ITS, community turnover was better explained by locality for this dataset. In the ITS data, we found that campinas and igapós are dissimilar in their communities (Fig. 4A). This can be explained by the physicochemical soil properties (Fig. 3). When it comes to chemical properties, campinas and igapós were placed at opposite extremes of PC1 and PC2 (Fig. 3B). With respect to the physical properties, campinas have plots in both extremes of PC1, but igapós were better explained by clay content (Fig. 3A). Clay content was an important factor in explaining leaf litter in central Amazon fungi56. For the 18S and COI, the similarity between habitats is better explained by comparing seasonally flooded and non-flooded habitats (Fig. 4). In both communities, igapós and várzeas are similar to each other and distinct from terra-firme and campinas. This is in agreement with results from studies of micro-organisms in general in the same areas10. These results were expected as the flooded period is a powerful factor that selects for a very specific vegetation type 66, 67, 68,26. Igapós and várzeas are more restricted to a fine soil texture, while in terra-firme and campinas the soil texture varies more (Fig. 3A). However, regarding the chemical properties, terra-firme and campinas have almost exclusively poor soils, while igapós and várzeas present different gradients of soil fertility (Fig. 3B). These distinct patterns among markers might be explained by the differences in taxonomic coverage of each marker, since different species of fungi have distinct habitat preferences12.
We were surprised to find that Campinas were, on average, the richest habitat for fungi. This stands in contrast to patterns observed for animals and plants69, 70, and fungi in Colombian Amazonia14. One explanation for the campinas being the richest environment may be the need for plants to associate with micro-organisms that fix nutrients in the poor soil habitats. For instance, some studies of campinas in Amazonia address the diversity of ectomycorrhizal fungi71, 72, 73, 74, 75. The general pattern is that the diversity of ectomycorrhizal fungal diversity is the highest in temperate zones12,76,77, but due to the poor soil in Campinas, the ectomycorrhizal fungi will be more diverse.
The origin of the campinas in Amazonia is a debated topic70, but the nature of their soil, caused by high drainage and high acidity, is considered one of the poorest in the world78. In this context, Singer et al.73 hypothesized that the ectomycorrhizal fungi increase the ability of their host plant to acquire nutrients and water in these very stressful habitats. We found a high richness and number of indicator OTUs in campinas (Fig. 5, Table 3), adding more evidence that the campinas may be hotspots for the diversity of fungi and other micro-organisms. However, we registered very few ectomycorrhizal indicator OTUs, although these results could be biased by the lack of representative DNA sequences from tropical areas79. It is interesting that várzea areas have on average 43 to 53% fewer OTUs of known mycorrhizal species for the three markers. Of the four habitats analysed, várzea soils are of higher fertility as they are flooded by nutrient-rich waters, decreasing the necessity for plants to associate with mycorrhizal fungi, in accordance with the hypothesis proposed by Singer et al.73.