Warming effects on microbial functional composition and C-decomposing genes. (A) the percentage of variation in microbial functional compositions explained by warming in the early cool season (January to February), warm season (March to September), and late cool season (October to December), as tested by Adonis. Significances are indicated by ** when P < 0.010 and * when P < 0.050. (B) Dissimilarities of microbial functional compositions between warming and control on a monthly basis. The dissimilarity values of paired warmed and control samples were fitted to nonlinear quadratic regression. The R2 and P-values were calculated, reflecting the variance explained by the regression. Distances were calculated based on the Bray-Curtis metric. (C) Response ratios showing changes in the abundance of functional genes associated with C decomposition between warmed and control samples in the early cool season (January to February), warm season (March to September), and late cool season (October to December). C substrates are arranged in the order from labile C to recalcitrant C. Error bars indicate 95% confidence intervals of abundance differences between warmed and control samples. (D) Percentages of significant changes of C-decomposing gene probes by warming. Probes are classified into four categories: unique under warming (probes detected only in warmed samples, and those that were likely present in control samples but below the level of detection), increased under warming (the response ratio > 0, P < 0.050), increased under warming (the response ratio > 0, P < 0.050), and unique under control (probes detected only in control samples, and those that were likely present in warmed samples but below the level of detection).
