Small-scale soil heterogeneity affects the distribution of arbuscular mycorrhizal fungal species in a hot-spot field in a Mediterranean site

Arbuscular mycorrhizal fungi (AMF) establish mutualistic symbioses with the roots of most plant species, playing a key role in crop productivity and ecosystem functioning. The characterization of local AMF genotypes assemblages is a prerequisite for the reproduction, both ex situ and on farm, of efficient native AMF communities, in order to improve soil quality and fertility through agricultural management and to use selected AMF as inoculants in sustainable agriculture. Contrasting evidences have been found by studies addressing the influence of different environmental drivers, among which geographical distance, on the diversity and composition of AMF communities in agricultural soils. In the present study, a Mediterranean "hot-spot" of AMF species richness was studied to assess whether AMF display a spatial structure at a local scale and whether their spore distribution is affected by soil variables. AMF distribution patterns were estimated by spore analysis, in 12 field plots, which were characterized for their chemical and physical variables. Plots were distributed along a transect across the field, and sampling points distances ranged from 4 to 102 m. While most taxa displayed an aggregated distribution, no significant spatial correlation was detected among AMF communities. On the contrary, AMF communities were significantly correlated with the entire set of soil variables, and with distinct soil characteristics, such as pH and total N, but not P. Among the most abundant taxa, Funneliformis geosporus, Septoglomus constrictum, Funneliformis mosseae, Ambispora granatensis and Glomus badium, were positively affected by total N, and Paraglomus occultum negatively affected by pH. Multiple regression analysis revealed that distribution of most AMF species was significantly related to nitrogen, clay, cation exchange capacity and pH. Multivariate (RDA) analysis confirmed results of multiple regression, with the best model selecting three soil variables (N, clay and pH). Variance partitioning analysis also showed that AMF distribution was mostly explained by environmental variables (36.4%) and by environmental variables spatial structure (16.4%). Overall, the results of this work revealed the importance of soil variables in shaping AMF community composition and spatial heterogeneity, within a site characterized by high AMF richness.