Determining functional traits that affect fungal niches for unobservable and unculturable taxa belowground, such as arbuscular mycorrhizal (AM) fungi, remains one of the largest challenges in microbial ecology. One promising avenue to elucidate AM fungal niches is to survey their distributions in natural ecosystems and estimate niche characteristics based on which conditions promote growth of particular AM fungal taxa. We used the MaarjAM database to assess plant host specificity by examining the phylogenetic signal of AM fungi colonizing 683 plant species and 420 AM fungal species. Using the same database, we also built species distribution models that predict relative occurrence rates of a focal AM fungal taxon given environmental conditions at locations where they are known to occur. These distribution models can then elucidate both overall range size and environmental conditions most strongly associated with AM fungal distributions at the global scale. We hypothesized that if AM fungi are constrained to specific plant hosts, their distributions should mimic host distributions. Otherwise, we expected that AM fungal distributions would largely reflect environmental factors (i.e., soil resources or climate). We further predicted that dispersal limitation would affect AM fungal distributions as a functional of spore morphology; taxa with larger spores would have more restricted distributions.
Results/Conclusions
Determinants of AM fungal distributions and ecological niches varied across taxa. Overall, 24% of AM fungi had phylogenetically clustered host associations and therefore distributions that could be potentially plant host-mediated. For other AM fungal taxa climate was a main determinant of AM fungal niches, with distributions constrained by the minimum annual temperature, minimum precipitation, and soil moisture. Some AM fungal taxa were influenced by soil pH. Yet few AM fungal species distributions were associated with global datasets of soil nutrient concentrations. Future sampling in underrepresented locations (e.g., the Southern hemisphere) and biomes (e.g., deserts) will improve estimates of environmental drivers of AM fungal distributions and niches. In addition, more AM fungal-plant association data is needed, as undersampling can bias results towards host specificity. Nevertheless, if AM fungi are more specific to particular plant hosts than expected based on historical paradigms, and are constrained by adverse temperature and moisture conditions, this warrants further consideration in our examination of both current and future AM fungal distributions.