Microbes play central roles in soil nutrient cycling, yet a limited range of
microbial community characteristics have been used to explain ecosystem nutrient
cycling rates and their importance relative to plant and abiotic factors remains
unclear. In this study, we assessed which of 126 commonly measured soil fungal
and bacterial community characteristics best explained soil nitrogen (N) and
phosphorus (P) cycling rates in temperate forests in the Northeastern U.S., as
well as the relative contributions of microbial, plant, and abiotic factors.
Using boosted regression tree modeling, we identified the microbial variables
with the highest contributions to models explaining nutrient cycling rates: the
relative abundances of ectomycorrhizal fungi and N-decomposition genes from
oligotrophic bacteria were the most important for net ammonification, the
relative abundances of indicator taxa in bacterial networks, nitrifying
bacteria, and copiotrophic bacteria were the most important for net
nitrification, and the relative abundance of fungal P-cycling oxidoreductase
genes was the most important for net soil phosphate change. Microbial variables
explained more variation than plant and abiotic variables in multivariate linear
models of net nitrification and net phosphate release rates, but not net
ammonification rates, which were largely explained by soil edaphic factors. Leaf
litter traits were also important in explaining variation in net nitrification
rates, and soil temperature was important in explaining rates of net phosphate
release in soil. Collectively, our findings suggest that the N-cycling capacity
of microbial functional guilds and fungal community P-cycling capacity should be
incorporated into ecosystem biogeochemical models to improve our predictions and
understanding of nutrient cycling and related ecological processes.