In dryland ecosystems, ecological processes are often explained by a pulse-dynamics framework, in which episodic precipitation events moisten dry soils, triggering brief pulses of biological activity as a result of increased resource availability. In this hierarchical framework, small infrequent rainfall events stimulate microbially-driven biogeochemical processes, while larger, or more tightly clustered small events are needed to stimulate larger biological response pulses like plant growth. In the southwestern United States, rainfall event size and frequency have become increasingly more variable over the past century. While pulse-dynamics research has traditionally focused on the response of aboveground plant communities, evidence suggests soil microbial communities are also sensitive to changes in rainfall variability. Yet, few studies have examined the temporal dynamics of microbial diversity and relative abundance in the context of the pulse-dynamics framework. To gain a better understanding of how changes in rainfall variability affect microbial communities through time, we performed 16S rRNA gene sequencing of rhizosphere soils in a rainfall manipulation experiment located in a northern Chihuahuan Desert grassland associated with the Sevilleta Long Term Ecological Research program. Soils were collected monthly during the summer monsoon (July-September) over a period of six days in conjunction with experimentally manipulated irrigation treatments that varied in both size and frequency (small: 5 mm weekly events (n=12); large: 20 mm monthly events (n=3)).
Results/Conclusions
We found that the microbial community was dominated by Proteobacteria (35.3–45.6%), Actinobacteria (17.4–26.5%), Acidobacteria (7.6–13.5%), and Bacteriodetes (8.3–11.2%), regardless of rainfall treatment or month. Alpha diversity, as measured by Chao1 richness, ranged from 274 in the July control treatment to 538 in the July small treatment. Microbial community composition, as determined by perMANOVA, was significantly affected by treatment (p=0.001) and month (p=0.001), but not their interaction. When compared to control plots, the composition of the microbial community was more significantly affected by the small treatment (p=0.001) than by the large treatment (p=0.020). Across the monsoon season, the difference in community composition between July and September was more significant (p=0.001) than differences between July and August (p=0.004) or August and September (p=0.017). While the composition of the dominant phyla was surprisingly consistent throughout the monsoon season, our results indicate changes in rainfall variability do affect these microbial communities through time. This knowledge will enable the development of more accurate conceptual models of dryland ecosystem structure and function, particularly in the context of a changing climate.