Plant Functional Traits Define Microbial Response to Nutrient Availability in Tropical Rainforest Soil

ABSTRACT High global inputs of nitrogen (N) compared with relatively low inputs of phosphorus (P) increase nutrient imbalances that may cause substantial shifts in plant functional traits and modulate resource utilization strategies, which are associated with soil microbial communities. These community‐level trait‐based adaptations and the responses of soil microbiomes to the projected nutrient changes remain largely unexplored. Here, we characterized the nutrient‐induced shifts in plant functional traits and microbial communities in P‐limited tropical rainforest soils by combining spatial multivariate analyses across 160 km 2 of primary and secondary tropical rainforest with an in situ 14‐year nutrient addition experiment. The links between plant traits and microbial composition depending on soil N and P contents were examined to test how vegetation regulates the responses of microbial communities to nutrient input. Elevated soil N increased P limitation and thus led to a shift in leaf traits representing a conservative economy, as indicated by increases in leaf N:P ratios and leaf dry matter content. In response to the conservative shift in plant traits, soil bacterial r‐strategists, arbuscular mycorrhizal and saprotrophic fungal guilds increased in relative abundance and thus were consistently enriched with increasing N content in soil. Addition of P to soil, however, led to increases in vegetation traits for acquisition economy, characterized by increases in leaf P content, specific leaf area, and trait diversity. With the shift to traits for acquisition in high‐P soils, the relative abundance of bacterial K‐strategists and ectomycorrhizal fungi rasied. Thus, vegetation traits have selective effects on soil microbiomes to acquire specific functions needed for P acquisition in P deficient tropical soil, which may, in turn, accelerate nutrient cycles and impact soil carbon sequestration. Our results suggest that models need to incorporate plant traits in predicting microbial dynamics and the associated functions under changing nutrient conditions.