Climate change is causing increased precipitation variability across ecosystems. The size and timing of rainfall events directly and indirectly impact ecosystem functions such as primary productivity. For California Mediterranean grasslands, we hypothesized that rainfall timing impacts plant production both 1) directly, because peak-season spring rainfall affects water availability during peak growth (e.g., greater aboveground production in high-rainfall springs) and peak-season drought limits this growth and causes earlier senescence, and 2) indirectly, where early-season fall precipitation patterns impact plant functional and trait compositional shifts. For example, after fall germination, frequent precipitation is associated with competitive, annual grasses, whereas fall drought is associated with a higher abundance of drought-tolerant forbs. We expect a continuous drought throughout the growing season to have both these direct and indirect effects. At the Sierra Foothills Research and Extension Center in Browns Valley, California, we experimentally tested how plant communities and their ecosystem functions shift across rainfall conditions using rainout shelters in which we manipulated rainfall amount and timing (early-season fall drought, peak-season spring drought, continuous drought and ambient rainfall) over three years and assessed plant responses: aboveground net primary production (ANPP), percent cover of species, diversity, phenology, and traits.
Results/Conclusions
Total ANPP was highest when rainfall occurred during the early-season fall (e.g., ambient rainfall or peak-season spring drought) and was lowest when rainfall occurred either primarily during peak-season (e.g., early-season fall drought) or under continuous drought. Specifically, early-season fall and consistent drought treatments negatively affected grass ANPP, while there were no significant effects of drought treatments on forb ANPP. Although total ANPP was not affected by peak-season spring drought, senescence occurred earlier in peak-season spring and consistent drought plots compared to ambient and early-season fall drought plots. Treatments did not affect percent cover of forbs or grasses, indicating no overall functional shift in the plant community with drought. However, species composition did shift with treatments. Continuous drought communities shifted toward species with stress-response traits such as greater coarse root diameter and higher proportion of fine roots. Overall, our results show that drought effects in California Mediterranean grasslands are season-dependent, where early-season fall drought indirectly affects total productivity via impacts on grasses, but peak-season spring drought directly impacts phenology, with earlier peak biomass and senescence.