Predictors of Arbuscular Mycorrhizal Fungal Communities in the Brazilian Tropical Dry Forest.

Arbuscular mycorrhizal fungi (AMF) are symbiotic fungi with a broad distribution, and many taxa have physiological and ecological adaptations to specific environments, including semiarid ecosystems. Our aim was to address regional distribution patterns of AMF communities in such semiarid environments based on spore morphological techniques. We assessed AMF spores at the bottom and top of inselbergs distributed throughout the tropical dry forest in the Northeast region of Brazil. Across 10 replicate inselbergs and the surrounding area, spanning a range of altitude between 140 and 2000 m, we scored the AMF soil diversity and properties in 52 plots. We fitted parsimonious ordination analyses and variance partitioning models to determine the environmental factors which explained the variation in AMF community, based on morphological spore analysis. The diversity of AMF was similar at the bottom and top of inselbergs; however, we detected high variation in abundance and richness across sites. We formulated a parsimonious richness model that used physical soil factors as predictors. The AMF community structure could be best explained through the variables coarse and total sand, iron, organic matter, potassium, silt, and sodium which together accounted for 17.8% of total variance. Several AMF species were indicators of either deficiency or high values of specific soil properties. We demonstrated that habitat isolation of the inselbergs compared with surrounding areas did not trigger differences in AMF communities in semiarid regions of Brazil. At the regional scale, soil predictors across sites drove the distribution of symbiotic mycorrhizal fungi.