While altered precipitation regimes can greatly impact biodiversity and ecosystem functioning, we lack a comprehensive view of how these impacts are mediated by changes to the seasonality of precipitation (i.e., whether it rains more/less in one season relative to another). Over two years we examined how altered seasonal precipitation influenced annual plant biomass and species richness, Simpson’s diversity, and community composition of annual plant communities in a dryland ecosystem that receives both winter and summer rainfall and has distinct annual plant communities in each season. Using a rainfall exclusion, collection, and distribution system we excluded precipitation and added water during each season individually and compared responses to control plots which received ambient summer and winter precipitation. In control plots, we found five times greater annual plant biomass, twice as many species, and higher diversity in winter relative to summer. Adding water increased annual plant biomass in summer only, did not change richness or diversity in either summer or winter, and modestly shifted community composition. Excluding precipitation in either season reduced annual plant biomass, richness and Simpson’s diversity. However, in the second winter season, biomass was higher in the plots where precipitation was excluded in the previous summer seasons suggesting that reduced productivity in the summer may facilitate biomass in the winter. Our results suggest that increased precipitation in summer may have stronger short-term impacts on annual plant biodiversity and ecosystem function relative to increased winter precipitation. In contrast, decreasing precipitation may have ubiquitous negative effects on annual plants across both summer and winter but may lead to increased biomass in the following off-seasons. These patterns suggest that annual plant communities exhibit asymmetries in their community and ecosystem responses to altered seasonal precipitation and that considering the seasonality of precipitation is important for predicting the effects of altered precipitation regimes.

To sample species composition and biomass of the annual plant community in each treatment, we established six 1-m² subplots in each plot. Within the central 0.5 m² of each subplot we measured community composition, in the same location every year, using a point-intercept approach with 50 points recording all hits at each point. In the area between where species composition was sampled and the outer edge of the sub-plot, we sampled aboveground biomass in an additional 0.25 m² area. We rotated the location of biomass sampling to a different side each season to not re-sample the same location, and for each harvest we clipped all aboveground biomass, dried it for 4-days at 60°C, and weighed it to the nearest 0.01 g. All subplots were placed at least 1m from the edge of the plot to reduce the potential for edge effects from the rainfall manipulation shelters.