Symbiotic associations between plant roots and soil fungi have played a major role in the evolution of land plants and continue to influence the dynamics of terrestrial ecosystems. The plant generally provides carbohydrates to fungal partners in exchange for limiting soil nutrients. The consequences for biogeochemical cycles, including key services like soil C storage, depend on the lineages involved. The vast majority of plants (>70%) form arbuscular mycorrhizal (AM) associations with glomeromycete fungi in predominantly equatorial habitats with relatively little soil C. Other plants that are prevalent in nutrient-rich polar habitats either exclude mycorrhizae altogether (NM: 12%) or form facultative arbuscular mycorrhizal associations (AMNM: 10%). A few ecologically dominant lineages (2%) form ectomycorrhizal associations (EM) with diverse dikarya in boreal and temperate forests that store much more C belowground. Variation in soil C storage and other ecosystem services may directly reflect distinctive dynamics of alternative mycorrhizal symbioses. Alternatively, biogeochemical differences may reflect other traits that covary with phylogeny. Using a dataset of more than 32.3 million occurrence records for over 4000 plant species, we assessed the relative roles of hypothesized climatic and soil nutrient drivers on transitions between mycorrhizal associations with models that explicitly accounted for plant phylogeny.
Results/Conclusions
Transitions between mycorrhizal states depended on temperature, soil P and plant phylogeny. Higher temperature favored the AM state over the NM state with facultative AMNM state favored at intermediate temperatures. Higher soil P favored the NM over AMNM, and AMNM over AM states. The strong effects of temperature and soil P were consistent with expectations based on both biogeographic patterns and functional studies on the stability of the AM mutualism. Furthermore, the magnitude, direction and phylogenetic signal for these effects were similar across major seed plant lineages. In contrast, none of the hypothesized predictors explained transitions involving the EM state because the extremely slow rate of evolution to and from this state produced such strong phylogenetic signal that it overwhelmed apparently strong ecological effects in poorly-supported models that ignored phylogeny. This result suggests that biogeochemical differences between AM and EM dominated forests, including important services like soil C storage, may be more parsimoniously explained by idiosyncrasies of a few peculiar plant lineages than inherent properties of the EM association. Overall, our results emphasize how incorporating phylogeny can bring power and precision to analyses of ecosystem services.