Predicting Microbial Functional Diversity for Decomposition along an Aridity Gradient

Soil microbial communities are rarely represented in soil models or with extreme simplifications due to their complexity. Acknowledging that temperature and moisture are the primary controls over microbial functional diversity, this research aims to determine the extent to which soil functional diversity can be predicted based on these factors. We used the aridity index (AI), as this easy-to-measure metric includes temperature and moisture. Following the YAS framework, a widely accepted trait-based approach to characterize soil microbial communities, we hypothesized that under an identical food source, the functional strategies employed by the community will go from high growth yield (Y) in humid areas to higher investment in stress tolerance (S) in arid areas. We also expected a trade-off between investment in S and Y, while relative investment in A (resource acquisition) should remain constant. We further hypothesized that AI is a decent predictor of the microbial investments into the Y, A, and S traits. We used the DEMENTpy model, an in silico simulator, to derive YAS investments for hypothetical soil microbial communities at five sites along an aridity gradient in Spain. We validated model simulations using mass loss from Rooibos tea samples from each site and employed a Dirichlet regression model to predict YAS investments, using AI. Contrary to the hypotheses, increasing aridity changes community investment from Y to A, with limited changes in S. The A strategy could be predicted considerably well based on AI, while Y and S could not. Together with further validation of our modeling results with experimental data, our findings lay the groundwork in deriving simple mathematical formulations that can be integrated into Earth system models, allowing for upscaling from genomes to Earth system processes.