Field experiment design for simulated warming and altered precipitation, qSIP incubation, and the growth responses of soil bacteria to changing climate regimes.

A To examine the effects of warming and altered precipitation on an alpine grassland ecosystem, two levels of temperature (T0, T+), and three levels of precipitation (-P, nP, +P) were established in 2011. The soil samples were collected in 2020 and used for 18O-qSIP incubation. B Potential interaction types between multiple climate factors. The diagram shows that two factors (X and Y) of warming and altered precipitation impact a biological response in the same direction (upper) or have opposing effects on when acting separately. Their combined effect could be additive, i.e. the sum of the two factor effects. Alternatively, the interaction types can be antagonistic or synergistic. Null model (we use the additive expectation as the null model here) provides the threshold for distinguishing between these interactions.

Species-specific shifts of 18O excess atom fraction (EAF-18O).

Bars represent 95% confidence intervals (CIs) of OTUs. Each circle represents an OTU and color indicates phylum. The open circles with gray bars represent OTUs with 95% CI intersected with zero (indicating no significant 18O enrichment); Closed circles represent the OTUs enriched 18O significantly, whose 95% CIs were away from zero (i.e., the OTUs had detectable growth).

Bacterial growth responses to climate change and the interaction types between warming and altered precipitation.

A The growth rates, and B responses of soil bacteria to warming and altered precipitation at the whole community level. C The growth rates, and D responses of the dominant bacterial phyla had similar trends with that of the whole community.

The growth responses and phylogenetic relationship of incorporators subjected to different interaction types under two climate scenarios.

A The phylogenetic tree of all incorporators observed in the grassland soils. The inner heatmap represents the single and combined factor effects of climate factors on species growth, by comparing with the growth rates in T0nP. SiE: single-factor effects; CoE: combined effects. The outer heatmap represents the interaction types between warming and altered precipitation under two climate change scenarios. B The proportions of positive or negative responses in species growth to single and combined manipulation of climate factors. C The proportions of species growth influenced by different interaction types under two climate change scenarios.

Within-species shift in interaction types contributed to the variance of the whole community growth response under two climate scenarios.

A Venn plots showing the overlaps of incorporators, and their interaction types between two climate scenarios. B The phylogenetic relationship of the 215 incorporators whose growth dynamics were influenced by the weak antagonistic interaction of warming × drought and by the neutralizing effect of warming × wet. The blue-green bars represent the average growth rates of incorporators across different climate treatments. The heatmap displayed the potential functions associated with carbon and nutrient cycles predicted by Picrust2. The values of function potential were standardized (range: 0-1). The “anta” represents the antagonistic interaction; “W×D” represents warming × drought and “W×W” represents warming × wet.

The higher level of antagonism of wet × warming than that of drought × warming.

A The antagonistic intensities were assigned to the five interaction types on a 5-point scale, from −1 to 3 for synergistic, additive, weak antagonistic, strong antagonistic and nullifying effect, respectively. B The overall antagonistic intensities of all incorporators under warming × drought and warming × wet scenarios were estimated by weighting the relative proportions of incorporators subjected to different interaction types.

The nearest taxon index (NTI) for incorporators subjected to different interaction types under two climate change scenarios.

P-value based on comparison of phylogenetic distance observed and that based on 1,000 permutations of a null model. Values in bold: P < 0.05. Low P-values (and positive indices) indicate that species are more phylogenetically related than expected by chance (phylogenetically clustering).